You are on page 1of 739


& Memory
John H. Byrne
University of Texas Medical School at Houston

Howard Eichenbaum
Boston University Laboratory of Cognitive Neurobiology

Henry L. Roediger, III

Washington University Department of Psychology

Richard F. Thompson
University of Southern California Program in Neuroscience


Frank Menchaca, Vice President

Jill Lectka, Director of Publishing Operations

The Macmillan Psychology Reference Series is an on-going collec-
tion of single-volume overviews of the disciplines of psychology.
Each volume in the series presents a comprehensive and accessible
snapshot of a psychology field for general readers, high school sen-
iors, and college freshmen. Child Development (2001), the first vol-
ume in the series, examines the contemporary issues in the field of
child and adolescent development. Each work will feature the latest
research in the field along with numerous real-world applications
and examples.
& Memory

John H. Byrne, Editor in Chief

Learning and Memory, Second Edition

2003 by Macmillan Reference USA. ALL RIGHTS RESERVED Cover images reproduced by permission of
Macmillan Reference USA is an imprint of No part of this work covered by the copy- William T. Greenough and James D. Churchill
The Gale Group, Inc., a division of Thomson right hereon may be reproduced or used in (synapses), Photodisc (MRI profile), and Corbis
Learning, Inc. any form or by any meansgraphic, electron- (brain).
ic, or mechanical, including photocopying,
Macmillan Reference USA and Thomson recording, taping, Web distribution, or infor- Since this page cannot legibly accommo-
Learning are trademarks used herein under mation storage retrieval systemswithout date all copyright notices, the acknowledg-
license. the written permission of the publisher. ments constitute an extension of the copy-
right notice.
For more information, contact For permission to use material from this
Macmillan Reference USA product, submit your request via Web at
300 Park Avenue South, or you
New York, NY 10010 may download our Permissions Request form
Or you can visit our Internet site at and submit your request by fax or mail to:
Permissions Department
The Gale Group, Inc.
27500 Drake Rd.
Farmington Hills, MI 48331-3535
Permissions Hotline:
248-699-8006 or 800-877-4253, ext. 8006
Fax: 248-699-8074 or 800-762-4058


Learning & Memory / edited by John H. Byrne. - - 2nd ed.

p. cm. - - (Macmillan psychology reference series)
Previous ed. published as: Encyclopedia of learning and memory. New
York : Macmillan, c1992.
ISBN 0-02-865619-9
1. Learning, Psychology of- -Encyclopedias. 2. Memory- -
Encyclopedias. 3. Learning in animals- -Encyclopedias. 4. Animal
memory- -Encyclopedias. 5. Neuropsychology- -Encyclopedias. I. Title:
Learning and memory. II. Byrne, John H. III. Series.

BF318 .E53 2002

153 . 103- -dc21 2002008357

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

List of Articles ..............................................................................ix
List of Contributors.......................................................................xv

Learning and Memory, Second Edition ..........................................1


Ken Wachsberger

Denise Evans
Gretchen Gordon
Bill Kaufman
Gina Misiroglu
Patricia Onorato
Angela Pilchak
Gregory Teague


Dean Dauphinais
Mary Grimes
Lezlie Light

Cindy Baldwin

Lori Hines

Evi Seoud

Rita Wimberley

Linda K. Fetters


Learning refers to the process of acquiring new infor- Despite the great progress that has been made,
mation, whereas memory refers to the retention of that many questions remain. Why will we remember exact-
information so that it can be retrieved at a later time. ly where we were on September 11, 2001 for the rest
The topics of learning and memory have intrigued of our lives, but we cannot remember what we had for
philosophers and writers for centuries, and for good lunch on this day two weeks ago? Why are we unable
reason. Learning and memory are so central to our to recall events from our infancy? Are there any drugs
daily lives that disruption in these functions can inter- that can improve our memory? What is Alzheimers
rupt our most routine activities. For example, consid- disease and how can it be treated? Learning and
er the extraordinary memory processes that must take Memory, Second Edition contains articles on these key
place for you to successfully drive from home to work. aspects of memory. In general, its goal is to provide
When you wake up, you need to recall whether or not readers with a comprehensive overview of key topics
that day is a workday. When you enter your car, you in learning and memory through brief articles written
need to remember the best route to work and any traf- by selected experts in the field.
fic information that you may have heard to expedite All of the entries are original contributions from
your arrival. Next, you have to recall how to drive your scholars and researchers, written for a readership of
car and the complex rules of the road. When you students, teachers, journalists, and members of the
arrive at work, you need to remember the names of educated public. The articles range in length from
your colleagues and draw upon a memory for job- 800 to 3,000 words. Entries are arranged alphabetical-
related vocabulary and common past experiences. ly; each is accompanied by a bibliography listing sug-
Most importantly, you need to remember the skills gestions for further reading. The entries are also
that allow you to perform your job. linked by a comprehensive set of cross-references that
The centrality of learning and memory to our appear at the end of many of the entries. For exam-
daily lives has led to intense analysis by psychologists ple, in the entry on infantile amnesia, one finds cross-
and neurobiologists for the past century, and it will references to the following entries: CHILDHOOD,
undoubtedly remain at the forefront of research DEVELOPMENT OF MEMORY IN; CODING PROCESSES:
throughout this new century as well. Learning and ORGANIZATION OF MEMORY; EPISODIC MEMORY; EXPERTS
memory systems are vast and diverse, yet scientists MEMORIES; GUIDE TO THE ANATOMY OF THE BRAIN;
have determined that memory can be divided into NATURAL SETTINGS, MEMORY IN; PROSE RETENTION.
two major types: memory for skills and habits, and In addition, blind entries facilitate access to main
memory for facts and events. Moreover, significant articles (for example, Drugs. See: COGNITIVE
progress has been made in understanding the parts of ENHANCERS; DRUGS AND MEMORY; ELECTROCONVULSIVE
the brain involved in learning and memory as well as THERAPY AND MEMORY LOSS; PHARMACOLOGICAL
in acquiring a basic understanding of the genes, pro- TREATMENT OF MEMORY DEFICITS.). These blind
teins, and signaling molecules that mediate memory entries are arranged alphabetically throughout
acquisition and storage. Learning and Memory, Second Learning and Memory and provide alternate access
Edition contains articles that discuss these findings. It points to direct the reader to appropriate articles An
also contains biographical sketches of some of the key additional feature is the use of guideposts or series
individuals, now deceased, who have contributed to introductions. Where there is a group of related arti-
the current understanding of learning and memory. cles on a single topic, a short guidepost is available to


orient the reader to the topic. For example, for the tion would not have been possible without the
area of invertebrate learning, a short introduction tremendous work of the current Editorial Board, who
presents an overview of the six entries that follow. identified the topics and their authors, and reviewed
This second edition of Learning and Memory builds each contribution. Special thanks also go to Jill
upon the success of the first, which was called Lectka, Director of Publishing Operations at
Encyclopedia of Learning and Memory when it was pub- Macmillan Reference USA, and Project Editor Ken
lished by Macmillan in 1992. We are indebted to Larry Wachsberger for supporting the concept of a second
Squire, the previous editor, for establishing the over- edition as well as ensuring that the production sched-
all direction and organization of the Encyclopedia, ule was maintained.
which this second edition maintains. This second edi- John H. Byrne


Learning: Behavioral Phenomena UNDERLYING LEARNING Robert S. Lockhart
F. Robert Brush Thomas J. Carew ORGANIZATION OF MEMORY

Activity-Dependent Regulation of MOLECULAR BASIS OF LONG-TERM Roger W. Schvaneveldt

Neurotransmitter Synthesis Cognitive Enhancers
Jack C. Waymire Gregg A. Phares
Gregor y M. Rose
John H. Byrne
Aging and Memory in Animals Collective Memory
Diana S. Woodruff-Pak Aristotle (384322 B.C.E.)
James V. Wertsch
Patricia Smith Churchland
Aging and Memory in Humans Comparative Cognition
Fergus I. M. Craik Attention and Memory Anthony A. Wright
Harold Pashler
Algorithms, Learning Concepts and Categories,
Andrew G. Barto Autobiographical Memory Learning of
Martin A. Conway Edward Wisniewski
Alzheimers Disease
BEHAVIORAL ASPECTS Bartlett, Frederic (18861969) Conditioning, Cellular and
Marilyn S. Albert David J. Murray Network Schemes for Higher-
ENGINEERED ANIMAL MODELS Philip N. Hineline Dean V. Buonomano
Philip C. Wong
Behavior Therapy Conditioning, Classical and
Donald L. Price
Stanley J. Rachman Instrumental
Amnesia, Functional I. Gormezano
John F. Kihlstrom Birdsong Learning
Gregor y F. Ball Declarative Memory
Daniel L. Schacter
Jacques Balthazart Neal Cohen
Amnesia, Infantile
Children, Development of Dj Vu
Harlene Hayne
Memory in Alan S. Brown
Amnesia, Organic
Robert V. Kail Dementia
Kelly Sullivan Giovanello
Meghan Saweikis Cynthia A. Munro
Mieke Verfaellie
Classical Conditioning: Discrimination and
Amnesia, Transient Global
Behavioral Phenomena Generalization
Mark Kritchevsky
John T. Green John Moore
Aplysia Joseph E. Steinmetz
Distributed Practice Effects
Coding Processes Robert L. Greene
Fred D. Lorenzetti Drugs and Memory
John H. Byrne Marc Marschark Gregor y M. Rose


Early Experience and Learning Glutamate Receptors and Their Hunter, Walter S. (18891954)
Seymour Levine Characterization Norman E. Spear
Deborah Suchecki M. N. Waxham
Hypnosis and Memory
Ebbinghaus, Hermann Guide to the Anatomy of the John F. Kihlstrom
(18501909) Brain
Implicit Memory
Endel Tulving OVERVIEW
Neil W. Mulligan
Yuri Koutcherov
Eccles, John (19031997)
George Paxinos Imprinting
Per O. Andersen
AMYGDALA Johan J. Bolhuis
Eidetic Imagery Alexander J. McDonald G. Horn
Robert G. Crowder BASAL FOREBRAIN Individual Differences in
Electroconvulsive Therapy and M-Marsel Mesulam Learning and Memory
Memory Loss BASAL GANGLIA Colin M. MacLeod
Larr y R. Squire Ann M. Graybiel
Infancy, Memory in
Emotion, Mood, and Memory Andrew N. Meltzoff
Paula Hertel Michael Mauk
Leslie J. Carver
Episodic Memory Helen Barbas Insect Learning
Endel Tulving Stewart Hendr y Brian H. Smith
Evolution and Learning HIPPOCAMPUS AND Charles I. Abramson
Intelligence and Memory
Menno P. Witter Anna T. Cianciolo
Experts Memories
NEURON Robert J. Sternberg
K. Anders Ericsson
Peter Somogyi
False Memories Interference and Forgetting
Kathleen B. McDermott Robert A. Bjork
Michael E. Hasselmo
Jason C. K. Chan PERIRHINAL CORTEX AND Invertebrate Learning
Learning in Humans Rebecca D. Burwell MEMORY PROCESSING IN BEES
Bennett G. Galef, Jr. SYNAPSE Randolf Menzel
Paul Rozin Kristen M. Harris ASSOCIATIVE LEARNING IN

Guthrie, Edwin R. (18861959) HERMISSENDA

Sara J. Shettleworth Ernest R. Hilgard Terr y J. Crow
Forgetting Habituation and Sensitization in
Alan Gelperin
Ian Neath Vertebrates
Michael Davis
Freud, Sigmund (18561939) M. David Egger Stephan Steidl
Matthew Hugh Erdelyi Catharine H. Rankin
Harlow, Harry F. (19051981)
Br yan E. Kolb

Morris Moscovitch Head Injury William N. Frost

Gordon Winocur Gerri Hanten NEUROGENETICS OF MEMORY IN
Genetic Substrates of Memory
Hebb, Donald (19041985) Tim Tully
Peter M. Milner James, William (18421910)
Fred J. Helmstetter
CEREBELLUM Hormones and Memory Sheldon White
Shaowen Bao James L. McGaugh Kamins Blocking Effect:
Lu Chen Benno Roozendaal
Neuronal Substrates
HIPPOCAMPUS Hull, Clark L. (18841952) Jeansok J. Kim
Alcino J. Silva Abram Amsel Mark G. Baxter

Knowledge Systems and Material- Mathematical Learning Theory Multiple-Memory Systems

Specific Memory Deficits W. K. Estes Mark G. Packard
Elkhonon Goldberg
McGeoch, John A. (18971942) Natural Settings, Memory in
William B. Barr
Robert G. Crowder Ulric Neisser
Konorski, Jerzy (19031973)
Measurement of Memory Neocortical Plasticity
Anthony Dickinson
Robert S. Lockhart
Language Learning: Humans
Memory Consolidation: ADAPTATION AND
Peter W. Culicover
Molecular and Cellular Processes REORGANIZATION
Language Learning: Nonhuman Alcino J. Silva Daniel J. Felleman
Memory Consolidation: AUDITORY CORTEX
Duane M. Rumbaugh
Prolonged Process of Norman M. Weinberger
E. Sue Savage-Rumbaugh
Jared P. Taglialatela
Larr y R. Squire SYSTEM
Lashley, Karl (18901958) Paul G. Shinkman
Memory Search
Richard F. Thompson
Barbara Anne Dosher MOTOR CORTEX
Learned Helplessness Brian McElree Jeffrey A. Kleim
Steven E. Maier
Learning Disabilities Michael J. Watkins Peter W. Hickmott
R. Holly Fitch
Mental Retardation (Intellectual Neural Computation
Learning Theory: A History Disabilities) APPROACHES TO LEARNING
Robert C. Bolles Robert M. Hodapp Terrence J. Sejnowski
Learning Theory: Current Status Metacognition about Memory CEREBELLUM
Steve Maren Thomas O. Nelson Masao Ito
Localization of Memory Traces
David J. Krupa Metaplasticity Bruce L. McNaughton
Thomas J. Carew NEOCORTEX
Long-Term Depression in the
Thomas M. Fischer Leif H. Finkel
Cerebellum, Hippocampus, and
Neocortex Migration, Navigation, and OLFACTORY CORTEX AS A MODEL
Verner P. Bingman Richard H. Granger, Jr.
Long-Term Potentiation
OVERVIEW: COOPERATIVITY AND Mnemonic Devices Neural Substrates of Avoidance
ASSOCIATIVITY Mark A. McDaniel Learning
Bengt Gustafsson Mnemonists Michael Gabriel
Eric Hanse Charles P. Thompson Neural Substrates of Classical
Modality Effects Conditioning
Thomas H.Brown
Derick H. Lindquist John M. Gardiner CARDIOVASCULAR RESPONSES
Nelson Cowan Donald A. Powell
Howard Eichenbaum Morphological Basis of Learning DISCRETE BEHAVIORAL RESPONSES

MAINTENANCE and Memory Joseph E. Steinmetz

Gar y Lynch Len Clear y Bill P. Godsil
J. David Sweatt James D. Churchill Michael Davis
Lorenz, Konrad (19031989) Motor Skill Learning Neural Substrates of Emotional
David F. Sherr y James C. Houk Memory
Luria, A .R. (19021977) Mller, Georg Elias (18501934) Glenn E. Schafe
Elkhonon Goldberg Hilde A.E. Lechner Joseph E. LeDoux

Neurogenesis Prefrontal Cortex and Memory in Savant Syndrome

Gerd Kempermann Primates Darold A. Treffert
Fred H. Gage Joaquin M. Fuster
Schizophrenia and Memory
Neuroimaging Primates, Visual Attention in Stephan Heckers
Steven E. Petersen Bharathi Jagadeesh Anthony P. Weiss
Neurotransmitter Systems and Primates, Visual Perception and School Learning
Memory Memory in Nonhuman M. C. Wittrock
Michel Baudr y Elisabeth A. Murray
Joel L. Davis Second Messenger Systems
Procedural Learning James H. Schwartz
Object Concept, Development of ANIMALS
Jeanne L. Shinskey Semantic Memory
Mark G. Packard
Observational Learning HUMANS

Indre V. Viskontas Claude G. Cech
Albert Bandura
Olds, James (19221973) Sharon L. Thompson-Schill
John F. Disterhoft Prose Retention
Mark A. McDaniel Semon, Richard (18591918)
Olton, David (19431994) Daniel L. Schacter
Matthew L. Shapiro Protein Synthesis in Long-Term
Memory in Vertebrates Sensory Memory
Operant Behavior
Steven P. R. Rose Nelson Cowan
W. K. Honig
Brent Alsop Ramn y Cajal, Santiago Serial Organization
(18521934) Alice F. Healy
Oral Traditions
Larr y W. Swanson
David C. Rubin Sex Differences in Learning
Reconstructive Memory Jocelyne B. Bachevalier
Orienting Reflex Habituation
Daniel M. Bernstein
E. N. Sokolov Skinner, B. F. (19041990)
Elizabeth F. Loftus
Oscillations, Synchrony, and Kurt Salzinger
Neuronal Codes Rehabilitation of Memory
Disorders Sleep and Memory Consolidation
Wolf Singer Robert Stickgold
Elizabeth L. Glisky
Parallel Distributed Processing Social Memory Processes
Models of Memory Reinforcement
Andrew G. Barto Barbara H. Basden
James L. McClelland
Reinforcement or Reward in Sometimes Opponent Process
Passive (Inhibitory) Avoidance,
Learning (SOP) Model, in Conditioning
Fear Learning
Susan E. Brandon
Almira Vazdarjanova
Richard H. Thompson Allan R. Wagner
Pavlov, Ivan (18491936)
CEREBELLUM Source Monitoring
John J. Furedy
Rodney A. Swain Marcia K. Johnson
Pharmacological Treatment of ELECTRICAL SELF-STIMULATION, Karen J. Mitchell
Memory Deficits BRAIN
Gregor y M. Rose Spatial Learning: Animals
Andreas Arvanitogiannis
Thomas S. Collett
Stanley J. Rachman Wolfram Schultz Spatial Memory
Timothy P. McNamara
Piaget, Jean (18961980) Repetition and Learning
Howard E. Gruber Robert L. Greene Spence, Kenneth (19071967)
Howard H. Kendler
Place Cells Retrieval Processes in Memory
Edvard I. Moser Henr y L. Roediger III Spinal Plasticity
Michelle L. Meade Michael M. Patterson
Place versus Response Learning
Revisited in the Brain Ribot, Thodule (18391916) Stress and Memory
Mark G. Packard Jean-Louis Signoret Tracey J. Shors

Taste Aversion and Preference Underwood, Benton (19151994) Working Memory

Learning in Animals Eugene B. Zechmeister ANIMALS
Diego E. Berman Anthony A. Wright
Vestibulo-Ocular Reflex (VOR)
Thorndike, Edward (18741949) Plasticity Humans
Gilbert Gottlieb Masao Ito Todd S. Braver
Tip-of-the-Tongue Phenomenon
Janet M. Duchek Visual Memory, Brightness and
Jessica M. Logan Flux in
David Lavond
Tolman, Edward C. (18861959)
Mark R. Rosenzweig Watson, John B. (18781958)
Donald A. Riley Eliot Hearst

Charles I. Abramson Shaowen Bao Diego E. Berman

Oklahoma State University University of California, San Weizmann Institute of Science,
Insect Learning Francisco Israel
Genetic Substrates of Taste Aversion and
Marilyn S. Albert
Memory: Cerebellum Preference Learning in
Harvard Medical School
Alzheimers Disease: Helen Barbas
Behavioral Aspects Daniel M. Bernstein
Boston University
Brent Alsop University of Washington, Seattle
Guide to the Anatomy of the
University of Otago, New Reconstructive Memory
Brain: Cerebral Cortex
Zealand Verner P. Bingman
Operant Behavior William B. Barr
Bowling Green State University
New York University School of
Abram Amsel Migration, Navigation, and
Medicine Homing
University of Texas Knowledge Systems and
Hull, Clark L. (18841952) Robert A. Bjork
Material-Specific Memory
Per O. Andersen Deficits UCLA
University of Oslo, Norway Interference and Forgetting
Eccles, John (19031997) Andrew G. Barto
Johan J. Bolhuis
University of Massachussetts
Andreas Arvanitogiannis Utrecht University, The
Algorithms, Learning
Concordia University, Montreal Netherlands
Reinforcement or Reward in Imprinting
Learning: Electrical Self- Barbara H. Basden Robert C. Bolles
Stimulation, Brain California State University, University of Washington (ret.)
Fresno Learning Theory: A History
Jocelyne B. Bachevalier
University of Texas Medical Social Memory Processes
Susan E. Brandon
School at Houston Michel Baudry Yale University and American
Sex Differences in Learning University of Southern Psychological Association
Gregory F. Ball California Sometimes Opponent
Johns Hopkins University Long-Term Potentiation: Process (SOP) Model, in
Birdsong Learning Maintenance Conditioning
Jacques Balthazart Neurotransmitter Systems Todd S. Braver
University of Lige, Belgium and Memory Washington University, St. Louis
Birdsong Learning Mark G. Baxter Working Memory: Humans
Albert Bandura Harvard University Alan S. Brown
Stanford University Kamins Blocking Effect: Southern Methodist University
Observational Learning Neuronal Substrates Dj Vu


Thomas H. Brown Genetic Substrates of McGeoch, John A.

Yale University Memory: Cerebellum (18971942)
Long-Term Potentiation:
James D. Churchill Peter W. Culicover
University of Illinois at Urbana- Ohio State University
F. Robert Brush Champaign Language Learning: Humans
Purdue University Morphological Basis of
Active and Passive Avoidance Learning and Memory: Joel L. Davis
Learning: Behavioral Vertebrates Office of Naval Research
Phenomena Neurotransmitter Systems
Patricia Smith Churchland and Memory
Dean V. Buonomano University of California, San
UCLA Diego Michael Davis
Conditioning, Cellular and Aristotle (384322 B.C.E.) Emory University
Network Schemes for Habituation and Sensitization
Anna T. Cianciolo in Vertebrates
Higher-Order Features of
Yale University Neural Substrates of Classical
Intelligence and Memory Conditioning: Fear-
Rebecca D. Burwell Potentiated Startle
Len Cleary
Brown University
University of Texas Medical Anthony Dickinson
Guide to the Anatomy of the
School at Houston University of Cambridge, UK
Brain: Perirhinal Cortex
Morphological Basis of Konorski, Jerzy (19031973)
and Associated Cortical
Learning and Memory:
Areas John F. Disterhoft
John H. Byrne Northwestern University Medical
Neal Cohen
University of Texas Medical School
University of Illinois
School at Houston Olds, James (19221973)
Declarative Memory
Aplysia: Classical
Barbara Anne Dosher
Conditioning and Operant Thomas S. Collett
University of California, Irvine
Conditioning University of Sussex, Brighton,
Memory Search
Aplysia: Molecular Basis of UK
Long-Term Sensitization Spatial Learning: Animals Janet M. Duchek
Martin A. Conway Washington University School of
Thomas J. Carew
University of Durham, UK Medicine, St. Louis
University of California, Irvine
Aplysia: Development of Autobiographical Memory Tip-of-the-Tongue
Processes Underlying Phenomenon
Nelson Cowan
Learning M. David Egger
University of Missouri
Metaplasticity University of Medicine and
Modality Effects
Leslie J. Carver Sensory Memory Dentistry of New Jersey
University of California, Habituation and Sensitization
Fergus I. M. Craik in Vertebrates
San Diego
Rotman Research Institute of
Infancy, Memory in Howard Eichenbaum
Baycrest Centre, Toronto

Claude G. Cech Aging and Memory in Boston University
University of Louisiana at Humans Long-Term Potentiation:
Lafayette Behavioral
Terry J. Crow
Semantic Memory: Cognitive Matthew Hugh Erdelyi
University of Texas Medical
Effects School at Houston Brooklyn College and the
Jason C. K. Chan Invertebrate Learning: Graduate School, City
Washington University, St. Louis Associative Learning in University of New York
False Memories Hermissenda Freud, Sigmund (18561939)
Lu Chen Robert G. Crowder K. Anders Ericsson
University of California, San Yale University (ret.) Florida State University
Francisco Eidetic Imagery Experts Memories

W. K. Estes Bennett G. Galef, Jr. Ann M. Graybiel

Indiana University McMaster University, Ontario Massachusetts Institute of
Mathematical Learning Food Aversion and Technology
Theory Preference Learning in Guide to the Anatomy of the
Humans Brain: Basal Ganglia
Michael S. Fanselow
UCLA John M. Gardiner John T. Green
University of Sussex, UK Indiana University
Neural Substrates of Classical
Modality Effects Classical Conditioning:
Conditioning: Fear
Behavioral Phenomena
Conditioning, Freezing Alan Gelperin
Monell Chemical Senses Center, Robert L. Greene
Daniel J. Felleman
Philadelphia Case Western Reserve University
University of Texas Medical Distributed Practice Effects
Invertebrate Learning:
School at Houston Associative Learning in Repetition and Learning
Neocortical Plasticity: Adult Limax
Visual CortexAdaptation William T. Greenough
Kelly Sullivan Giovanello University of Illinois at Urbana-
and Reorganization
Boston University School of Champaign
Leif H. Finkel Medicine Morphological Basis of
University of Pennsylvania Amnesia, Organic Learning and Memory:
Neural Computation: Vertebrates
Elizabeth L. Glisky
Neocortex Howard E. Gruber
University of Arizona
Thomas M. Fischer Rehabilitation of Memory Teachers College, Columbia
Wayne State University Disorders University (ret.)
Piaget, Jean (18961980)
Metaplasticity Bill P. Godsil
UCLA Bengt Gustafsson
R. Holly Fitch
Neural Substrates of Classical Gteborg University, Sweden
University of Connecticut, Storrs
Conditioning: Fear Long-Term Potentiation:
Learning Disabilities
Conditioning, Freezing Overview: Cooperativity and
William N. Frost Associativity
Elkhonon Goldberg
Chicago Medical School Eric Hanse
New York University School of
Invertebrate Learning: Gteborg University, Sweden
Habituation and Long-Term Potentiation:
Knowledge Systems and
Sensitization in Tritonia Material-Specific Memory Overview: Cooperativity and
Deficits Associativity
John J. Furedy
University of Toronto Luria, A .R. (19021977) Gerri Hanten
Pavlov, Ivan (18491936) I. Gormezano Baylor College of Medicine,
University of Iowa (ret.) Houston
Joaquin M. Fuster Head Injury
UCLA Conditioning, Classical and
Instrumental Kristen M. Harris
Prefrontal Cortex and
Gilbert Gottlieb Boston University
Memory in
University of North Carolina at Guide to the Anatomy of the
Michael Gabriel Brain: Synapse
Chapel Hill
University of Illinois and Thorndike, Edward Michael E. Hasselmo
Beckman Institute (18741949) Boston University
Neural Substrates of Guide to the Anatomy of the
Avoidance Learning Richard H. Granger, Jr.
Brain: Olfactory Cortex
University of California, Irvine
Fred H. Gage Neural Computation: Harlene Hayne
Salk Institute for Biological Olfactory Cortex as a Model University of Otago, New
Studies, La Jolla, CA for Telencephalic Zealand
Neurogenesis Processing Amnesia, Infantile

Alice F. Healy Masao Ito Yuri Koutcherov

University of Colorado Brain Science Institute, RIKEN, Prince of Wales Medical Research
Serial Organization Japan Institute, Australia
Long-Term Depression in the Guide to the Anatomy of the
Eliot Hearst
Cerebellum, Hippocampus, Brain: Overview
University of Arizona
and Neocortex
Watson, John B. (18781958) Mark Kritchevsky
Neural Computation:
University of California, San
Stephan Heckers Cerebellum
Massachusetts General Hospital Vestibulo-Ocular Reflex Amnesia, Transient Global
Schizophrenia and Memory (VOR) Plasticity
David J. Krupa
Fred J. Helmstetter Bharathi Jagadeesh Duke University
University of Wisconsin, University of Washington, Seattle Localization of Memory
Milwaukee Primates, Visual Attention in Traces
Genetic Substrates of
Marcia K. Johnson David Lavond
Memory: Amygdala
Yale University University of Southern California
Stewart Hendry Source Monitoring Visual Memory, Brightness
Johns Hopkins University and Flux in
Robert V. Kail
Guide to the Anatomy of the
Purdue University Hilde A. E. Lechner
Brain: Cerebral Cortex
Children, Development of Salk Institute for Biological
Paula Hertel Memory in Studies, La Jolla, CA
Trinity University Mller, Georg Elias
Gerd Kempermann
Emotion, Mood, and (18501934)
Max Delbrck Center for
Molecular Medicine and Joseph E. LeDoux
Peter W. Hickmott Humboldt University, Berlin New York University
University of California, Neurogenesis Neural Substrates of
Riverside Emotional Memory
Howard H. Kendler
Neocortical Plasticity:
University of California, Santa Harvey Levin
Somatosensory Cortex
Barbara (ret.) Baylor College of Medicine,
Ernest R. Hilgard Spence, Kenneth Houston
Stanford University (ret.) (19071967) Head Injury
Guthrie, Edwin R.
John F. Kihlstrom Seymour Levine
University of California, Berkeley University of California, Davis
Philip N. Hineline Amnesia, Functional Early Experience and
Temple University Hypnosis and Memory Learning
Jeansok J. Kim Derick H. Lindquist
Robert M. Hodapp Yale University Yale University
UCLA Kamins Blocking Effect: Long-Term Potentiation:
Mental Retardation Neuronal Substrates Amygdala
(Intellectual Disabilities)
Jeffrey A. Kleim Robert S. Lockhart
W. K. Honig University of Lethbridge, Alberta, University of Toronto
Dalhousie University, Halifax, Canada Coding Processes: Levels of
Canada (ret.) Neocortical Plasticity: Motor Processing
Operant Behavior Cortex Measurement of Memory
G. Horn Barbara J. Knowlton Elizabeth F. Loftus
Cambridge University UCLA University of Washington, Seattle
Imprinting Procedural Learning: Reconstructive Memory
James C. Houk Jessica M. Logan
Northwestern University Medical Bryan E. Kolb Washington University, St. Louis
School University of Lethbridge, Canada Tip-of-the-Tongue
Motor Skill Learning Harlow, Harry F. (19051981) Phenomenon

Fred D. Lorenzetti James L. McGaugh Neil W. Mulligan

University of Texas Medical University of California, Irvine Southern Methodist University
School at Houston Hormones and Memory Implicit Memory
Aplysia: Classical
Timothy McNamara Cynthia A. Munro
Conditioning and Operant
Conditioning Vanderbilt University Johns Hopkins University School
Spatial Memory of Medicine
Gary Lynch Dementia
University of California, Irvine Bruce L. McNaughton
Long-Term Potentiation: University of Arizona David J. Murray
Maintenance Neural Computation: Queens University at Kingston,
Hippocampus Canada
Colin M. MacLeod Bartlett, Frederic
University of Toronto at Michelle L. Meade
Scarborough Washington University, St. Louis
Individual Differences in Retrieval Processes in Elisabeth A. Murray
Learning and Memory Memory National Institute of Mental
Steven E. Maier Andrew N. Meltzoff Primates, Visual Perception
University of Colorado University of Washington, Seattle and Memory in Nonhuman
Learned Helplessness Infancy, Memory in
Ian Neath
Steve Maren Randolf Menzel Purdue University
University of Michigan Freie Universitt Berlin Forgetting
Learning Theory: Current Invertebrate Learning:
Status Associative Learning and Ulric Neisser
Marc Marschark Memory Processing in Bees Cornell University
Rochester Institute of Technology Natural Settings, Memory in
M-Marsel Mesulam
Coding Processes: Imagery Thomas O. Nelson
Feinberg Medical School of
Michael Mauk Northwestern University University of Maryland, College
University of Texas Medical Guide to the Anatomy of the Park
School at Houston Brain: Basal Forebrain Metacognition about
Guide to the Anatomy of the Memory
Brain: Cerebellum Peter M. Milner
McGill University, Montreal Mark G. Packard
James L. McClelland Hebb, Donald (19041985) Yale University
Carnegie Mellon University Multiple-Memory Systems
Parallel Distributed Karen J. Mitchell Place versus Response
Processing Models of Yale University Learning Revisited in the
Memory Source Monitoring Brain
John Moore Procedural Learning:
Mark A. McDaniel
University of New Mexico University of Massachusetts- Animals
Mnemonic Devices Amherst Harold Pashler
Prose Retention Discrimination and University of California, San
Generalization Diego
Kathleen B. McDermott
Washington University, St. Louis Morris Moscovitch Attention and Memory
False Memories University of Toronto and Michael M. Patterson
Alexander J. McDonald Rotman Research Institute, University of California, San
University of South Carolina Toronto Diego
School of Medicine Frontal Lobes and Episodic Spinal Plasticity
Guide to the Anatomy of the Memory
George Paxinos
Brain: Amygdala Edvard I. Moser Prince of Wales Medical Research
Brian McElree Norwegian University of Science Institute, Australia
New York University and Technology Guide to the Anatomy of the
Memory Search Place Cells Brain: Overview

Steven E. Petersen Mark R. Rosenzweig James H. Schwartz

Washington University, St. Louis University of California, Berkeley Columbia University
Neuroimaging Tolman, Edward C. Second Messenger Systems
Gregg A. Phares (18861959)
Terrence J. Sejnowski
University of Texas Medical Paul Rozin Salk Institute for Biological
School at Houston University of Pennsylvania Studies, San Diego
Aplysia: Molecular Basis of Food Aversion and Neural Computation:
Long-Term Sensitization Preference Learning in Approaches to Learning
Donald A. Powell Humans
Matthew L. Shapiro
Dorn VA Medical Center and
David C. Rubin Mount Sinai School of Medicine,
University of South Carolina
Duke University New York
Neural Substrates of Classical
Conditioning: Oral Traditions Olton, David (19431994)
Cardiovascular Responses Duane M. Rumbaugh David F. Sherry
Donald L. Price Georgia State University University of Western Ontario
Johns Hopkins University School Language Learning: Evolution and Learning
of Medicine Nonhuman Primates Lorenz, Konrad (19031989)
Alzheimers Disease: Human Kurt Salzinger Sara J. Shettleworth
Disease and the Genetically American Psychological University of Toronto
Engineered Animal Models
Association Foraging
Stanley J. Rachman Skinner, B. F. (19041990)
University of British Columbia Paul G. Shinkman
E. Sue Savage-Rumbaugh University of North Carolina at
Behavior Therapy
Phobias Georgia State University Chapel Hill
Language Learning: Neocortical Plasticity:
Catharine H. Rankin Nonhuman Primates Development of the Visual
University of British Columbia
Meghan Saweikis System
Invertebrate Learning: C. ele-
gans Purdue University Jeanne L. Shinskey
Children, Development of University of Denver
Donald A. Riley
Memory in Object Concept,
University of California, Berkeley
Tolman, Edward C. Daniel L. Schacter Development of
(18861959) Harvard University Tracey J. Shors
Henry L. Roediger III Amnesia, Functional Rutgers University, Picataway,
Washington University, St. Louis Semon, Richard (18591918) NJ
Retrieval Processes in Glenn E. Schafe Stress and Memory
Memory New York University Jean-Louis Signoret
Benno Roozendaal Neural Substrates of Salpetriere Hospital, Paris (ret.)
University of California, Irvine Emotional Memory Ribot, Thodule (18391916)
Hormones and Memory Petra Scheck Alcino J. Silva
Gregory M. Rose University of Maryland, College UCLA
Memory Pharmaceuticals Corp., Park Genetic Substrates of
Montvale, NJ Metacognition about Memory Memory: Hippocampus
Cognitive Enhancers
Wolfram Schultz Memory Consolidation:
Drugs and Memory
Pharmacological Treatment University of Cambridge Molecular and Cellular
of Memory Deficits Reinforcement or Reward in Processes
Learning: Striatum
Steven P. R. Rose Wolf Singer
Open University, UK Roger W. Schvaneveldt Max Planck Institute for Brain
Protein Synthesis in Long- Arizona State University, East Research, Frankfurt, Germany
Term Memory in Coding Processes: Oscillations, Synchrony, and
Vertebrates Organization of Memory Neuronal Codes

Brian H. Smith Rodney A. Swain Almira Vazdarjanova

Ohio State University University of Wisconsin- University of Arizona
Insect Learning Milwaukee Passive (Inhibitory)
Reinforcement or Reward in Avoidance, Fear Learning
E. N. Sokolov Learning: Cerebellum
Moscow Lomonosov State Mieke Verfaellie
Larry W. Swanson
University (ret.) Boston University School of
University of Southern California
Orienting Reflex Habituation Medicine and Boston VA
Ramn y Cajal, Santiago
Peter Somogyi (18521934) Healthcare System
University of Oxford, UK Amnesia, Organic
J. David Sweatt
Guide to the Anatomy of the Baylor College of Medicine, Indre V. Viskontas
Brain: Neuron Houston UCLA
Norman E. Spear Long-Term Potentiation: Procedural Learning:
Signal Transduction Humans
Binghamton University
Mechanisms and Early
Hunter, Walter S.
Events Allan R. Wagner
Jared P. Taglialatela Yale University
Larry R. Squire Georgia State University Sometimes Opponent
VA Medical Center, San Diego Language Learning: Process (SOP) Model, in
and University of California, Nonhuman Primates Conditioning
San Diego
Charles P. Thompson Michael J. Watkins
Electroconvulsive Therapy
Kansas State University Rice University
and Memory Loss
Mnemonists Memory Span
Memory Consolidation:
Prolonged Process of Richard F. Thompson M. N. Waxham
Reorganization University of Southern California
University of Texas Medical
Lashley, Karl (18901958)
Stephan Steidl School at Houston
Richard H. Thompson Glutamate Receptors and
University of British Columbia
University of British Columbia Their Characterization
Invertebrate Learning: C. ele-
Reinforcement or Reward in
gans Jack C. Waymire
Learning: Anatomical
Joseph E. Steinmetz Substrates University of Texas Medical
Indiana University Sharon L. Thompson-Schill School at Houston
Classical Conditioning: University of Pennsylvania Activity-Dependent
Behavioral Phenomena Semantic Memory: Regulation of
Neural Substrates of Classical Neurobiological Perspective Neurotransmitter Synthesis
Conditioning: Discrete
Darold A. Treffert Norman M. Weinberger
Behavioral Responses St. Agnes Hospital, WI University of California, Irvine
Robert J. Sternberg Savant Syndrome Neocortical Plasticity:
Yale University Tim Tully Auditory Cortex
Intelligence and Memory Cold Spring Harbor Laboratory,
NY Anthony P. Weiss
Robert Stickgold Massachusetts General Hospital
Invertebrate Learning:
Harvard Medical School Schizophrenia and Memory
Neurogenetics of Memory
Sleep and Memory
in Drosophila
Consolidation James V. Wertsch
Endel Tulving Washington University, St. Louis
Deborah Suchecki Rotman Research Institute of Collective Memory
Universidade Federal de So Baycrest Centre, Toronto
Paulo Ebbinghaus, Hermann Sheldon White
Early Experience and (18501909) Harvard University
Learning Episodic Memory James, William (18421910)

Gordon Winocur Brain: Hippocampus and Anthony A. Wright

Trent University, Peterborough, Parahippocampal Region University of Texas Medical
and Rotman Research Institute, School at Houston
M.C. Wittrock
Toronto Comparative Cognition
Frontal Lobes and Episodic Working Memory: Animals
School Learning
Memory Eugene B. Zechmeister
Philip C. Wong
Edward Wisniewski Loyola University
Johns Hopkins University School
University of North Carolina at Underwood, Benton
of Medicine
Greensboro (19151994)
Alzheimers Disease: Human
Concepts and Categories, Disease and the Genetically
Learning of Engineered Animal Models
Menno P. Witter Diana S. Woodruff-Pak
Vrije Universiteit Medical Center, Temple University
Amsterdam Aging and Memory in
Guide to the Anatomy of the Animals
ACTIVE AND PASSIVE AVOIDANCE the specified response during the WS-shock interval
LEARNING: BEHAVIORAL terminates the WS and prevents the occurrence of the
shock. In the passive form, suppression of the speci-
fied response during the WS-shock interval prevents
Avoidance learning is the behavioral product of an in- the occurrence of the shock. In both forms an inter-
strumental (operant) training procedure in which a trial interval (ITI) intervenes between successive pre-
predictable aversive event, typically electric shock, sentations of the WS, usually in the range of 0.5 to 5.0
does not occur contingent upon the occurrence or minutes.
nonoccurrence of a specified response by the learning
In the active free-operant procedure there are no
organism. Avoidance training occurs in two forms: ac-
discrete trials signaled by WSs. Instead, the avoidance
tive and passive. In the active form, the avoidance contingency is dependent on time. Specifically, two
contingency depends on the occurrence of a specified timers control events: a response-shock (R-S) timer
response on the part of the organism; in the passive (e.g., set for thirty seconds) and a shock-shock (S-S)
form, the avoidance contingency depends on the non- timer (e.g., set for five seconds). Training starts with
occurrence (i.e., the suppression) of some specified re- the S-S timer operating. Every time it runs out, it re-
sponse. The response to be suppressed may be either starts and delivers an inescapable shock of some dura-
spontaneous or learned by virtue of prior reward tion (e.g., 0.5 second). The specified response turns
training. In both forms, however, the avoidance con- off the S-S timer and starts the R-S timer. Every addi-
tingency consists of the prevention or omission of a tional response resets the R-S timer to its full value.
predictable noxious event. Noxious events are de- If the R-S timer runs out, it presents a shock and starts
fined in terms of the preference relation in which the the S-S timer (Sidman, 1953). This procedure has
absence of the event is preferred (measured by been used only in the active form. A variation of this
choice) to the presence of the event. Usually the nox- procedure eliminates the S-S timer and makes shock
ious event is electric shock, but loud noise, blasts of termination contingent upon the specified response
air, and high and low temperatures have been used. rather than upon a fixed duration of shock.
Avoidance training also utilizes one of two proce- In addition, a free-operant passive procedure
dures: discrete-trial or free-operant. In the discrete- known as punishment simply takes a response, which
trial procedure a distinctive stimulus, called a warning occurs spontaneously or by virtue of prior reward
signal (WS), signals the organism that the occurrence training, and makes shock or some other aversive
of the aversive event (e.g., electric shock) is imminent. event contingent on the occurrence of that response.
In most experiments the WS-shock interval is five to The response is usually suppressed. This is also called
sixty seconds in duration. In the active form, making passive-avoidance training. It has been used in a proce-


dure in which an animal such as a mouse or rat runs This theoretical problem was ostensibly solved by
from a brightly lighted elevated platform into a dark Mowrer (1950), supported by Solomon and Wynne
compartment where it receives a single electric shock. (1954) and Rescorla and Solomon (1967), by postulat-
The tendency to enter the dark compartment is in- ing that Pavlovian conditioning of fear on early es-
nate, and the single punishment results in subsequent cape trials, in which the WS is paired with shock, pro-
long latencies to reenter the dark compartment. This vided the acquired motivation to terminate the WS
is called one-trial passive-avoidance training, and it (now a conditioned aversive stimulus), thus providing
has been used extensively in the study of memory be- secondary (acquired) negative reinforcement for the
cause the learning event is fixed in time, which allows escape-from-fear response (i.e., the avoidance re-
analysis and manipulation of temporarily constrained sponse). Others thought that the fear response was in-
neuropharmacological and endocrine processes asso- strumentally reinforced by the termination of shock
ciated with learning. Alternatively, a hungry animal (Miller, 1951), but the upshot was the same: Reduc-
may initially be rewarded with food for pressing a tion of fear by termination of the WS, whether ac-
lever and subsequently shocked for making that same quired by Pavlovian or instrumental means, was the
response. Usually several shocks are required to sup- source of the acquired negative reinforcement for the
press the lever pressing. avoidance response. Thus, two processes were postu-
lated: acquisition of fear during escape trials (by Pav-
Warner (1932) was the first to use a discrete-trial
lovian or operant conditioning) and acquisition of the
active-avoidance procedure to study the association
span of the white rat (using WS-shock intervals of one instrumental avoidance response, reinforced by fear
to thirty seconds); he used what has become known as reduction. This theoretical interpretation was sup-
a shuttle box, a two-compartment box in which the ported by the results of an elegant experiment by
animal is required to run or jump back and forth be- Kamin (1956). Additional research in support of two-
tween the two compartments to avoid the shock. process theory used a transfer paradigm in which ani-
mals were given Pavlovian conditioning in one situa-
These procedures and the behaviors they pro- tion, and the effects of those conditioned stimuli were
duce have been of interest to psychologists since the observed when they were subsequently superimposed
early studies of behaviorism in the United States. on an operant baseline of responding in another situ-
John Watson and, especially, Edward Thorndike pos- ation (Solomon and Turner, 1960). This two-process
tulated that learned responses were a product of their theory provides the best account of avoidance learn-
consequences. That is, a response occurs, a pleasur- ing in its various forms.
able or aversive event ensues, and the response is re-
inforced (increases) if the event is pleasurable or pun- Some animals of most species learn the avoidance
ished (decreases) if the event is aversive. contingency, whether in the active or the passive
form, using discrete-trial or free-operant procedures.
Hilgard and Marquis (1940), two early behavioral Dogs are particularly adept at avoidance learning in
theorists, had trouble accounting for avoidance learn- an active, discrete-trial shuttle-box procedure and
ing because it was a product of a procedure where the typically show strong resistance to extinction (Solo-
reinforcing event was the response-contingent absence mon and Wynne, 1954). In contrast, rats are particu-
of an event, not the response-contingent presence of an larly difficult to train in an active, lever-press, dis-
event: crete-trial procedure and require special training
procedures (Berger and Brush, 1975). Thus, there
Learning in this [avoidance] situation ap- are important differences among species and re-
pears to be based in a real sense on the avoid-
sponse requirements. Additionally, in all forms of
ance of the shock. It differs clearly from other
avoidance learningactive and passive, discrete-trial
types of instrumental training in which the
and free-operantthere are enormous individual dif-
conditioned response is followed by a definite
ferences. Some individuals of whatever species learn
stimulus changefood or the cessation of
rapidly and well, whereas others do not (Brush,
shock [reward training (positive reinforce-
ment) or escape training (negative reinforce-
ment)]. In instrumental avoidance training In view of these findings it is not surprising that
the new response is strengthened in the ab- several investigators have genetically selected for dif-
sence of any such stimulus; indeed, it is ferences in avoidance learning. Bignami (1965) re-
strengthened because of the absence of such ported the first experiment with Wistar albino rats in
a stimulus. Absence of stimulation can obvi- which the selectively bred phenotypes were good or
ously have an influence on behavior only if poor at avoidance learning in a shuttle box. The re-
there exists some sort of preparation for or sulting strains are known as the Roman High Avoid-
expectation of the stimulation. (pp. 5859) ance and Roman Low Avoidance strains (RHA and

RLA, respectively). Training consisted of five daily avoidance learning is strongly influenced by genetic
sessions of fifty trials each. Selection was based on the factors, and many behavioral, physiological, and ana-
number of avoidance responses during the first two tomical correlates of avoidance learning have been
sessions (many or few) and on good or poor retention identified. Several of those correlates appear to be
from each session to the next. Selection was highly ef- closely linked, genetically, to the avoidance pheno-
fective because, by the fifth generation, the RHA and types. Researchers are trying to identify the mecha-
RLA animals avoided, respectively, on 68 percent and nisms by which genes determine avoidance learning.
20 percent of the trials. Modern molecular-genetic technology might enable
In 1977 Brush reported on the development of them to identify those genes.
the Syracuse High Avoidance and Syracuse Low
Avoidance strains (SHA and SLA, respectively). Long- See also: NEURAL SUBSTRATES OF AVOIDANCE
Evans hooded rats were trained for sixty trials in auto- LEARNING; PASSIVE (INHIBITORY) AVOIDANCE,
mated shuttle boxes. The data from over twenty gen- FEAR LEARNING
erations of selection indicated that shuttle-box avoid-
ance learning is heritable: SHA and SLA animals
avoided on 67 percent and 0 percent of the sixty trials
Bammer, G. (1978). Studies on two new strains of rats selectively
of training. Realized heritability (h2, which can range bred for high or low conditioned avoidance responding.
between 0.0 and 1.0; Falconer, 1960) was estimated Paper presented at the Annual Meeting of the Australian Soci-
to be 0.16 in each strain, a value comparable with that ety for the Study of Animal Behavior, Brisbane.
found in other selection studies (Brush, Froehlich, Berger, D. F., and Brush, F. R. (1975). Rapid acquisition of dis-
and Sakellaris, 1979). crete-trial lever-press avoidance: Effects of signal-shock inter-
val. Journal of the Experimental Analysis of Behavior 24, 227239.
In 1978 Bammer reported on the first six genera- Bignami, G. (1965). Selection for high rates and low rates of avoid-
tions of selective breeding of Sprague-Dawley albino ance conditioning in the rat. Animal Behavior 13, 221227.
rats for high and low levels of avoidance responding Brush, F. R. (1966). On the differences between animals that learn
and do not learn to avoid electric shock. Psychonomic Science
in a shuttle box. The resulting strains are known as
5, 123124.
the Australian High Avoidance and Australian Low (1977). Behavioral and endocrine characteristics of rats se-
Avoidance strains (AHA and ALA, respectively). lectively bred for good and poor avoidance behavior. Activitas
Training consisted of fifty trials in one or more daily Nervosa Superioris 19, 254255.
sessions. Realized heritability over the first five gener- Brush, F. R., Froehlich, J. C., and Sakellaris, P. C. (1979). Genetic
ations of selection was 0.18 and 0.27 for the AHA and selection for avoidance behavior in the rat. Behavior Genetics
9, 309316.
ALA strains, respectively. Falconer, D. S. (1960). Introduction to quantitative genetics. London:
A unidirectionally selected strain, known as the Oliver and Boyd.
Tokai High Avoider (THA), was bred in Japan from Hilgard, E. R., and Marquis, D. G. (1940). Conditioning and learning.
New York: Appleton-Century-Crofts.
Wistar stock using a lever-press response and a free- Kamin, L. J. (1956). The effect of termination of the CS and avoid-
operant procedure (S-S = 5 seconds, R-S = 30 sec- ance of the US on avoidance learning. Journal of Comparative
onds, shock duration = 0.5 second). The selection cri- and Physiological Psychology 49, 420424.
terion was an avoidance rate of more than ninety-five Miller, N. E. (1951). Learnable drives and rewards. In S. S. Stevens,
percent in the last five of ten daily one-hour training ed., Handbook of experimental psychology. New York: Wiley.
Mowrer, O. H. (1950). On the dual nature of learninga reinter-
sessions. Selection was successful: THA males and fe-
pretation of conditioning and problem solving. In Mo-
males learn faster and to a higher level of perfor- wrers Learning theory and personality dynamics. New York: Ron-
mance than unselected control animals from the orig- ald Press.
inal stock. Rescorla, R. A., and Solomon, R. L. (1967). Two-process learning
theory: Relationships between Pavlovian conditioning and in-
The fact that so many selective breeding experi- strumental learning. Psychological Review 74, 151182.
ments for avoidance behavior have been successful is Sidman, M. (1953). Two temporal parameters of the maintenance
a clear indicator of the extent to which this kind of be- of avoidance behavior by the white rat. Journal of Comparative
havior is under genetic control. In each experiment and Physiological Psychology 46, 253261.
the individual variability within each strain becomes Solomon, R. L., and Turner, L. H. (1960). Discriminative classical
conditioning under curare can later control discriminative
less as selection progresses, and it appears not to mat- avoidance responses in the normal state. Science 132, 1,499
ter what the details of the training procedures are. 1,500.
For example, SHA animals do better than controls in Solomon, R. L., and Wynne, L. C. (1954). Traumatic avoidance
a free-operant procedure, and THA animals do better learning: The principles of anxiety conservation and partial
than controls in discrete-trial, shuttle-box training. irreversibility. Psychological Review 61, 353385.
Warner, L. H. (1932). The association span of the white rat. Journal
Similarly, AHA animals outperform ALA animals in
of Genetic Psychology 39, 5789.
a discrete-trial avoidance task quite different from the
one in which they were selected. Thus, it is clear that F. Robert Brush

ACTIVITY-DEPENDENT REGULATION equilibrium with choline acetyltransferase (CAT), the

OF NEUROTRANSMITTER enzyme that catalyzes acetylcholine synthesis (Jope,
1979). Consequently, acetylcholine synthesis is well
below its maximal possible rate. Therefore any
Activity-dependent regulation of neurotransmitter synthesis change in the concentration of acetylcholine or its
refers to the ability of some nerve cells to change the precursors produces a change in acetylcholine syn-
amount of neurotransmitter synthesized in response thesis. This has been verified by the demonstration
to activity. Study of this regulation is prompted by the that the transport of choline into the cholinergic neu-
belief that it is important not only for maintaining a ron controls acetylcholine synthesis. For example,
source of neurotransmitter but also for adaptive choline addition to slices of brain tissue markedly in-
changes that take place in certain nerve cells during creases the acetylcholine synthesis rate. In addition,
learning and memory. A basic postulate necessary for increased neuronal activity increases choline uptake
neurotransmitter synthesis regulation to be a mecha- (Simon and Kuhar, 1975). This increased uptake oc-
nism for learning and memory is that increased curs in a manner that persists far beyond the period
neurotransmitter synthesis acts to increase neuro- of increased activity. Thus, stimulation of choline up-
transmitter secretion and, as a consequence, synaptic take is not due merely to a shift in the equilibrium of
strength. It has not been technically possible thus far the CAT-catalyzed reaction; instead, choline uptake
to demonstrate a causal relationship between activity- is regulated by neural activity per se. This regulation
dependent regulation of neurotransmitter synthesis of choline uptake by nerve activity is predicted to
and an increase in neurotransmitter secretion. None- maintain the strength of cholinergic synapses and is
theless, activity-dependent regulation of neurotran- a candidate to increase the capacity of these synapses
smitter synthesis remains a candidate for the cause of to secrete acetylcholine. Despite this evidence for ac-
neuroplastic changes that underlie learning, memo- tivity-dependent regulation of choline uptake, little is
ry, and neuroplasticity. Neurotransmitters that have known concerning how this regulation occurs. Thus,
shown activity-dependent regulation of their biosyn- the link between nerve activity and choline transport
thesis are acetylcholine, dopamine, and norepineph- remains unknown.
rine. This entry reviews the mechanisms of the syn- Catecholaminesdopamine, norepinephrine, and
thesis of these three neurotransmitters and possible epinephrineare neurotransmitters in the sympa-
roles of their regulatory mechanisms in learning and thetic limb of the autonomic nervous system and in
memory. several groups of neurons in the CNS. In contrast
Depending upon both the type of nerve cell and with the lack of a mechanism linking nerve activity
the time scale over which adaptation occurs, the cellu- and the regulation of choline uptake, catecholamine-
lar and biochemical mechanisms responsible for ac- synthesizing cells have several mechanisms in place to
tivity-dependent regulation of neurotransmitter syn- regulate catecholamine levels in response to nerve ac-
thesis vary. The time scale of changes ranges from tivity, both in peripheral and central nervous systems
very rapid changes (seconds), in which covalent modi- and at short- and long-term levels. As one review stat-
fication of enzyme protein structure is involved, to ed, An intricate scheme has evolved whereby tyro-
more delayed, longer-term changes (days). The latter sine hydroxylase activity is modulated by nearly every
involve alterations in genetic expression and turnover documented form of regulation (Kumer and Vrana,
of enzymes responsible for neurotransmitter biosyn- 1996). Regulation occurs at the step in catechol-
thesis. Catecholamines are regulated at both short- amines synthesis where tyrosine is hydroxylated to
and long-term levels, while acetylcholine is regulated form L-dopa. The enzyme catalyzing this reaction, ty-
only at a short-term level. The regulatory mecha- rosine hydroxylase, is the first of four enzymatic steps
nisms are often similar to cellular and biochemical in the catecholamine synthesis pathway. Because tyro-
mechanisms used in nonneural cells to regulate the sine hydroxylase is present in lower concentration
than the other enzymes of the synthesis pathway, it re-
synthesis of hormones or to regulate the biochemical
stricts the amount of neurotransmitter synthesized. In
pathways of intermediary metabolism.
addition, tyrosine hydroxylase is constitutively inhib-
Acetylcholine is the neurotransmitter in the auto- ited by the binding of a mole of catecholamine to each
nomic nervous system, the central nervous system mole of enzyme. Two interacting mechanisms are im-
(CNS), and at the neuromuscular junction. Even portant in short-term, nervous-activity regulation of
though acetylcholine has one of the highest rates of catecholamine synthesis. One mechanism is catechol-
turnover of any neurotransmitter, its concentration amine end-product inhibition of tyrosine hydroxylase
within nerve tissue fluctuates very little. This is be- activity by the catecholamine products of the pathway.
cause the precursors for acetylcholine synthesis, ace- The second is modification of tyrosine hydroxylase
tylcoenzyme A and choline, exist in a steady-state structure by the placement of phosphate groups on

the tyrosine hydroxylase molecule. The latter pro- Catecholamine synthesis is also regulated at a
cess, termed phosphorylation, is catalyzed by at least chronic, long-term level in response to persistent or
four separate protein kinases (Kumer and Vrana, extreme neural activation. In this case the amount of
1996), which phosphorylate tyrosine hydroxylase on the enzymes in the catecholamine synthetic pathway,
a combination of three serine residues in the N termi- and especially tyrosine hydroxylase, is elevated in re-
nal domain of the enzyme. The phosphorylation al- sponse to increased synaptic activity. For example,
ters the properties of tyrosine hydroxylase, increasing drugs or conditions such as stress that increase the au-
its catalytic activity. Phosphorylation is commonly tonomic nerve activity increase the level of tyrosine
used to modify proteins involved in regulation. hydroxylase in peripheral autonomic cells (Thoenen,
Mueller, and Axelrod, 1969). Because cutting the in-
The protein kinases that phosphorylate tyrosine nervation to these cells blocks the increase, the influ-
hydroxylase are cyclic adenosine monophosphate- ence is thought to be transynaptic. Because a rise in
dependent protein kinase (PKA), which phosphoryl- tyrosine hydroxylase mRNA precedes the increase in
ates serine 40; calcium-calmodulin-dependent pro- protein, the mechanism is believed to be increased
tein kinase II (CaM KII), which phosphorylates serine transcription of the mRNA encoding tyrosine hy-
19; and extracellular receptor activated protein ki- droxylase. The observation that drugs that increase
nase (ERK) (Haycock, Ahn, Cobb, and Krebs, 1992), CNS neuronal activity also induce increased levels of
which phosphorylates serine 31. This phosphoryl- CNS tyrosine hydroxylase shows that central tyrosine
ation has three major effects on the enzymes func- hydroxylase levels are also regulated by neuronal ac-
tion; all three changes enhance tyrosine hydroxylase tivity. This nerve activitydependent regulation of ty-
activity and increase catecholamine synthesis. Serine rosine hydroxylase synthesis is an attractive candidate
40 phosphorylation increases the affinity of the en- for learning and memory because the increase in ty-
zyme for tetrahydrobiopterin cofactor (the cofactor is rosine hydroxylase level is expected to increase the
normally present below optimal concentrations) and strength of the activated synapses. Whether or not
reduces the catecholamine binding and inhibition of this is the case is not yet known. Even so, a consider-
tyrosine hydroxylase (Daubner, Lauriano, Haycock, able effort is being carried out to understand as much
and Fitzpatrick, 1992). Serine 19 phosphorylation as possible about transynaptic regulation of tyrosine
causes the enzyme to interact with an activating pro- hydroxylase level because it is likely the best example
tein termed 14-3-3 protein (Ichimura et al., 1987; Ita- known of synaptically mediated regulation of protein
gaki et al., 1999). Serine 31 phosphorylation in- synthesis.
creases the catalytic activity by an as yet undefined Among the issues that remain unresolved con-
mechanism (Haycock, Ahn, Cobb, and Krebs, 1992). cerning the mechanism of the long-term regulation
of tyrosine hydroxylase is the nature of the intracellu-
Which of these mechanisms act to regulate the
lar mechanisms responsible for increased transcrip-
synthesis of catecholamines in response to neural ac-
tion. In a model tissue, the adrenal medulla, acetyl-
tivity, and how may they relate to learning and memo- choline and pituitary adenylyl cyclase activating
ry? Although these questions have been difficult to polypeptide (PACAP) are each able to modulate tyro-
answer, some conclusions are possible. Under circum- sine hydroxylase level. In this tissue either PACAP or
stances of cell depolarization, such as when a nerve acetylcholine stimulates cAMP level and activates
impulse invades the nerve terminal or during cholin- PKA. One hypothesis is that PKA migrates to the cell
ergic stimulation at the adrenal medulla, the phos- nucleus to regulate the rate of tyrosine hydroxylase
phorylation and activation of tyrosine hydroxylase by transcription by phosphorylating CREB (Cyclic AMP
Ca/CAM kinase II appear to be responsible for initial Response Element Binding), a regulatory protein as-
activation of tyrosine hydroxylase (Waymire et al., sociated with the tyrosine hydroxylase gene (Kuro-
1988; Waymire and Craviso, 1993) with a later activa- sawa, Guidotti, and Costa, 1976). Unresolved issues
tion through both PKA and ERK. In addition, evi- in this simple model are 1. whether acetylcholine
dence indicates that the phosphorylation on serine 19 stimulates a rise in cAMP or 2. whether other trans-
may facilitate the phosphorylation on serine 40 mitters, such as PACAP, are involved. Because
(Bevilaqua et al., 2001). Thus there appears to be a neuropeptides are secreted along with acetylcholine
time-dependent hierarchical phosphorylation and at some cholinergic synapses, it has been suggested
activation of tyrosine hydroxylase that occurs in steps that these agonists are responsible for long-term reg-
to activate catecholamine biosynthesis. Any condition ulation of tyrosine hydroxylase level (Wessels-Reiker,
that increases the size of these phosphorylation Haycock, Howlett, and Strong, 1991). Also, it is not
events is predicted to enhance the synthesis of cat- clear whether transcription is regulated solely
echolamines. This may translate into increased through PKA. Because several regulatory domains
neurotransmitter release and synaptic strength. exist in the tyrosine hydroxylase gene, it appears that

protein kinases other than the PKA are likely to be in- Haycock, J. W. (1990). Phosphorylation of tyrosine hydroxylase in
volved. In addition it is possible that the rapidly syn- situ at serine 8, 19, 31 and 40. Journal of Biological Chemistry
265, 11,68211,691.
thesized protein c-Fos may serve as a protein factor Haycock, J. W., Ahn, N. G., Cobb, M. H., and Krebs, E. G. (1992).
regulating tyrosine hydroxylase transcription. An ad- ERK1 and ERK2, two microtubule-associated protein 2 ki-
ditional question is whether increased transcription is nases, mediate the phosphorylation of tyrosine hydroxylase at
the principal mechanism of regulation. In isolated serine-31 in situ. Proceedings of the National Academy of Sciences
adrenal medullary chromaffin cells, transcription in- of the United States of America 89, 2,3652,369.
Ichimura, T., Isobe, T., Okuyama, T., Yamauchi, T. and Fujisawa,
creases for only a few hours following either acetyl- H. (1987). Brain 14-3-3 protein is an activator protein that ac-
choline of PACAP stimulation, whereas the increase tivates tryptophan 5-monooxygenase and tyrosine 3-
in mRNA occurs over several days and remains elevat- monooxygenase in the presence of Ca2+, calmodulin-
ed long after transcription has subsided. This raises dependent protein kinase II. FEBS Letters 219, 7982.
the issue of the stabilization of the tyrosine hydroxy- Itagaki, C., Isobe, T., Taoka, M., Natsume, T., Nomura, N., Hori-
gome, T., Omata, S., Ichinose, H., Nagatsu, T., Greene, L. A.,
lase mRNA as a major component of the synaptic reg- and Ichimura, T. (1999). Stimulus-coupled interaction of ty-
ulation. Indeed, tyrosine hydroxylase mRNA is capa- rosine hydroxylase with 14-3-3 proteins. Biochemical Journal
ble of regulation through stability, as has been 38, 15,67315,680.
demonstrated for its stabilization by elevated oxygen Jope, R. S. (1997). High affinity choline transport and acetylCoA
tension (Paulding and Czyzyk-Krzeska, 1999). production in brain and their roles in the regulation of acetyl-
choline synthesis. Brain Research Review 1, 313344.
The understanding of the activity-dependent Kumer, S. C., and Vrana, K. E. (1996). Intricate regulation of tyro-
regulation of catecholamines is much more complete sine hydroxylase activity and gene expression. Journal of
than for other neurotransmitters, such as acetylcho- Neurochemistry 67, 443462.
Kurosawa, A., Guidotti, A., and Costa, E. (1976). Induction of tyro-
line. For some neurotransmitter systemsthe amino sine 3-monooxygenase elicited by carbamylcholine in intact
acids and purines, for examplealmost nothing is and denervated adrenal medulla: Role of protein kinase acti-
known about their synthesis regulation, so it is not vation and translocation. Molecular Pharmacology 12, 420432.
clear whether it is activity-dependent. This is primari- Paulding, W. R., and Czyzyk-Krzeska, M. F. (1999). Regulation of
ly because these compounds are so intimately associ- tyrosine hydroxylase mRNA stability by protein-binding, py-
rimidine-rich sequence in the 3-untranslated region. Journal
ated with intermediary metabolism that it is difficult of Biological Chemistry 274, 2,5322,538.
to separate their neurotransmitter-related metabo- Simon, J. R., and Kuhar, M. J. (1975). Impulse-flow regulation of
lism from that associated with general cell function. high affinity choline uptake in brain cholinergic nerve termi-
One generalization emerging from the studies of the nals. Nature 255, 162163.
mechanisms of activity-dependent regulation of cat- Thoenen, H., Mueller, R. A., and Axelrod, J. (1969). Trans-
synaptic induction of adrenal tyrosine hydroxylase. Journal of
echolamine synthesis is the prominent position pro- Pharmacology and Experimental Therapeutics 169, 249254.
tein phosphorylation plays in both short- and long- Waymire, J. C., and Craviso, G. L. (1993). Multiple site phosphory-
term regulation. In the future, studies will likely be di- altion and activation of tyrosine hydroxylase. Advances in Pro-
rected to applying the understanding being gained of tein Phosphatases 7, 495506.
the mechanisms regulating catecholamine synthesis Waymire, J. C., Johnston, J. P., Hummer-Lickteig, K., Lloyd, A.,
Vigny, A., and Craviso, G. L. (1988). Phosphorylation of bo-
to other neurotransmitters. And although it is impor- vine adrenal chromaffin cell tyrosine hydroxylase: Temporal
tant to continue to investigate the mechanisms in- correlation of acetylcholines effect on site phosphorylation,
volved in activity-dependent regulation of neurotran- enzyme activation, and catecholamine synthesis. Journal of Bi-
smitter synthesis, it is also important to recognize that ological Chemistry 263, 12,43912,447.
the role of neurotransmitter synthesis regulation in Wessels-Reiker, M, Haycock, J. W., Howlett, A. C., and Strong, R.
(1991). Vasoactive intestinal polypeptide induces tyrosine hy-
higher functions, such as learning and memory, is still droxylase in PC12 cells. Journal of Biological Chemistry 266,
hypothetical. 9,3479,350.

Jack C. Waymire

Bevilaqua, L. R., Graham, M. E., Dunkley, P. R., Nagy-Felsobuki,
E. I., and Dickson, P. W. (2001). Phosphorylation of Ser(19) AGING
alters the conformation of tyrosine hydroxylase to increase See: AGING AND MEMORY IN ANIMALS; AGING
the rate of phosphorylation of Ser(40). Journal of Biological
Chemistry 276, 40,41140,416.
Daubner, S. C., Lauriano, C., Haycock, J. W., and Fitzpatrick, P.
F. (1992). Site-directed mutagenesis of serine 40 of rat tyro- DISEASE: HUMAN DISEASE AND THE
sine hydroxylase. Effects of dopamine and cAMP-dependent GENETICALLY ENGINEERED ANIMAL MODELS;
phosphorylation on enzyme activity. Journal of Biological Chem- PHARMACOLOGICAL TREATMENT OF
istry 267, 12,63912,646. MEMORY DEFICITS

AGING AND MEMORY IN ANIMALS conditioning is a simple form of learning that can be
studied with little modification across a variety of spe-
The passage of time produces changes in both the be- cies, including humans. There are at least four advan-
havior and the brains of organisms. A number of use- tages to using eyeblink conditioning as an animal
ful animal models of learning and memory in normal model of the effects of aging on learning and memory
aging have expanded the knowledge base and ex- in humans. First, both animals and humans show age-
tended the prospects for ameliorating learning and associated deficits in conditioning, and these can be
memory deficits. Completion of the mapping of the easily dissociated from age-associated changes in sen-
human and mouse genome and the development of sory systems (i.e., differences in CS thresholds) or
transgenic mouse models in the 1990s have accelerat- motor systems (differences in UR amplitude). Sec-
ed insights about mechanisms of learning, memory, ond, both animals and humans show age-associated
and aging. Since the mid-1990s, mouse models of changes in the neural substrates critical for eyeblink
neuropathology in Alzheimers disease have become conditioning, the cerebellum and the hippocampus.
available for behavioral testing. Two features of ani- Third, age-associated deficits have been artificially in-
mal models make them invaluable: First, the life duced with drugs in both young animals and young
spans of most animals are considerably shorter than humans. Finally, age-associated deficits can be re-
the human life span, compressing the time required versed in normal older animals using cognition-
to observe processes of aging. Second, invasive or enhancing drugs.
high-risk observations and experimental manipula-
tions are feasible with animals but not with humans. The critical substrate for eyeblink conditioning is
the cerebellum. The hippocampus plays a modulato-
Aging is most typically associated with declines in ry role in acquisition of CRs. Abnormal functioning
functioning, both neural and behavioral. However, of the hippocampus retards the rate of eyeblink con-
individual organisms age at different rates. One or- ditioning in the delay procedure. In addition, the
ganism may show a steady decline in functioning, hippocampus is essential for eyeblink classical condi-
whereas another shows only slight changes over the tioning procedures involving greater complexity,
years. An important goal toward an understanding of such as trace conditioning. In the trace procedure,
aging processes is to determine how changes in neu- the CS is presented and then turned off, and a blank
ral structures impact behavior. As such, behavioral period ensues before the onset of the US. The blank
paradigms that engage well-defined neural substrates period is called the trace and is shown in the left
are particularly valuable. Two such neurobiologically panel of Figure 1.
well-characterized paradigms will be highlighted:
eyeblink classical conditioning and spatial learning In eyeblink classical conditioning, intervals be-
and memory. tween the CS and US (called interstimulus intervals,
ISIs) of 250 and 750 milliseconds are the most com-
mon, with the most rapid conditioning between 250
Eyeblink Classical Conditioning and 500 milliseconds. The trace procedure extends
The basic classical conditioning procedure, the ISI in addition to inserting the blank trace period.
named the delay procedure by Ivan Pavlov, in- Thus, it increases difficulty in two ways. Holding the
volves repeated trials in which the presentation of an ISI constant at 750 milliseconds, researchers com-
initially neural stimulus, such as a tone (the condi- pared the delay and trace procedures in three-
tioned stimulus, or CS), is followed after approxi- month-old and twenty-four-month-old rabbits. There
mately half a second by a stimulus that evokes a re- were significant effects of age and procedure (see Fig-
flexive eyelid closure, such as a corneal air puff (the ure 1). Older rabbits performed more poorly in both
unconditioned stimulus, or US). The CS turns on and procedures, and both age groups performed more
remains on while the US is delivered and the two poorly in the trace than in the delay procedure. Com-
stimuli coterminate. Initially, eye blinks occur only parison of a number of studies testing the delay and
after the US, and the blinks are a reflexive response trace procedure in older rabbits indicated that age
(the unconditioned response, or UR). Eventually, eye differences in conditioning appeared earlier when
blinks occur after the CS but before the US. This is a the trace procedure was used.
learned response (the conditioned response, or CR). Cerebellar Substrates of Impaired
Thus, learning is defined as the acquisition of CRs. Conditioning in Older Animals
Richard F. Thompson suggested that the eye- Purkinje cells in the cerebellum integrate CS and
blink classical conditioning paradigm might be the US input and show patterns of engagement during
Rosetta Stone for brain substrates of age-related defi- eyeblink conditioning. In rabbits, Purkinje cell counts
cits in learning and memory (Thompson, 1988). have been carried out using histological techniques
Thompsons major point was that eyeblink classical after behavioral testing with eyeblink classical condi-

Figure 1

Left: Delay and trace procedures using same 750 milliseconds ISI (interstimulus intervals). Right: The number of trials it took rabbits
to attain a learning criterion of eight conditioned responses in nine consecutive trials. The higher the trials to criterion measure, the
slower the animals were to learn. A total of twenty-four young and seventeen old rabbits were tested in either the 750-milliseconds
delay or the 750-milliseconds trace (250 milliseconds CS, 500 milliseconds trace) procedure. There were statistically significant
differences between the younger and older rabbits, and there were statistically significant differences between the delay and trace
procedures. Older rabbits learned more slowly in both the delay and trace procedures, and both younger and older rabbits learned
more slowly in the trace than in the delay procedure.

tioning. The correlations between Purkinje cell num- Logan, and Thompson, 1987). Older rabbits were sig-
ber and eyeblink classical conditioning in rabbits were nificantly impaired in acquiring CRs. Furthermore,
high and statistically significant (Woodruff-Pak, older rabbits showed significantly less neural activity
Cronholm, and Sheffield, 1990). The fewer Purkinje in the US period than young rabbits by the second
cells a rabbit had, the longer it took it to acquire CRs. session of training. Matthew McEchron and John Dis-
Further analysis demonstrated that this relationship terhoft also reported that older rabbits had less hip-
was relatively independent of age because there was pocampal responsivity in the US period.
a highly significant correlation between Purkinje cell
number and conditioning when only young rabbits In a series of experiments in young and older rab-
were included. Mutant mice without Purkinje cells bits, Disterhoft and McEchron (2000) found that con-
condition slowly and produce fewer CRs, whereas ditioning-related hippocampal pyramidal-cell activity
their wild type littermates with Purkinje cells condi- varied across cells and that the different response
tion normally (Chen et al., 1996). Given the essential profiles were differentially affected by aging. Patterns
role of the cerebellum in the acquisition of CRs, Pur- of hippocampal pyramidal cell activation associated
kinje cell loss may account for a significant portion of with acquisition of trace eyeblink conditioning were
the age-related difference in eyeblink classical condi- different from activity recorded after CRs became as-
tioning. ymptotic. Pyramidal-cell activity associated with ac-
quisition was more sensitive to the effects of aging
Hippocampal Substrates of Impaired (McEchron, Weible, and Disterhoft, 2001). Various
Conditioning in Older Animals patterns of single-unit pyramidal-cell activity were
There is some evidence of age differences in hip- identified, and three response patterns were different
pocampal activity during eyeblink conditioning. Neu- between young and older rabbits that learned and
ronal responses were recorded from the dorsal hippo- those aged rabbits that did not. The patterns showed
campus of young (three-month-old), middle-aged significant changes during the first five days of condi-
(twenty-six-to-thirty-three-month-old), and older tioning in the young and aged learners, but the pat-
(thirty-nine-to-fifty-month-old) rabbits during eye- terns showed no change in the aged nonlearning
blink conditioning in the 750-millisecond trace pro- group. If these cells function to hold important infor-
cedure shown in Figure 1 (Woodruff-Pak, Lavond, mation for consolidation in other neural structures,

age-related deficits in conditioning may be ameliorat- trieve the correct map may explain why old rats are
ed by enhancing the function of these cells. more likely to make behavioral map-retrieval errors.
Older organisms, including rats, monkeys, and hu-
mans, have a greater tendency to become lost. Altered
Spatial Learning and Memory hippocampal plasticity mechanisms may underlie
Spatial behavioral tests evaluate the ability of the these changes in cognition that occur during normal
organism to know or to have a representation of its aging.
location in the environment and thus to navigate ef-
fectively. The intact functioning of the hippocampus Transgenic Mouse Models of Alzheimers
is necessary for learning and remembering spatial Disease
tasks. In old age, spatial memory is less efficient in hu-
mans and animals. Severe memory loss is the most prominent cogni-
tive symptom of Alzheimers disease (AD), and a fun-
When single cells are recorded in the hippocam- damental role in the pathogenesis of AD is brain de-
pus of a behaving rat, firing rate increases when the position of -amyloid (A). Mutations in the amyloid
animal is in a particular place in the environment. precursor protein (APP) and presenilin-1 (PS1) genes
These cells have been called place cells, and the are linked to forms of AD that are carried in families
area over which these cells show increased firing rates and called familial AD. By altering APP metabolism,
are called the cells place fields. Carol Barnes these mutations result in increased brain levels of A
(2001) described deficits that occur in the develop- peptide. Transgenic mice harboring mutant forms of
ment and maintenance of hippocampal place fields in the APP and/or PS1 gene associated with AD in hu-
old rats. As rats explore the environment, there is a mans are valid tools in the study of pathophysiologi-
change in the pattern of hippocampal place-cell dis- cal and behavioral effects of those genes in AD. Due
charge that occurs as a consequence of experience. to the relatively short life span of mice (two to three
For example, when rats run laps around a track in years), a high overexpression of the transgene is nec-
one direction, there is an expansion of the place fields essary to achieve the development of AD-like symp-
and a shift in their centers of mass toward the origin toms in the animals. The first transgenic mouse mod-
of the route. In old rats, there is a striking reduction els of AD were produced in the mid-1990s, and
in this experience-dependent form of plasticity in the thereafter there were a number of mouse models of
old place cells. Barnes suggested that the lack of field AD that developed A-containing plaques in the hip-
broadening observed in old rats might be expected to pocampus and neocortex, thus modeling human AD.
lead to a loss of precision in the information transmit- The hippocampus is engaged in eyeblink classical
ted as a consequence of experience. Ability to remem- conditioning and in spatial learning and memory,
ber routes may also be impaired by this deficit in and both of these behaviors are profoundly impaired
place-cell field broadening. in AD. Spatial learning and memory is the behavior
Barnes also found that the same memory- most commonly tested in transgenic mouse models of
impaired old rats that showed deficits in experience- AD, although eyeblink classical conditioning is a use-
dependent place-field expansion also retrieved inap- ful alternative behavioral measure that has direct par-
propriate hippocampal maps on some occasions. allels and can be tested in humans diagnosed with
Even when these memory-impaired old rats were in AD. A frequently used behavioral test for AD mouse
familiar environments, they retrieved inappropriate models is the Morris water maze, in which mice are
hippocampal maps from time to time. When a young placed in a pool of water that is opaque (to hide a plat-
rat is exposed to a familiar environment on one day form) and must learn the location of that platform to
and then exposed to that environment again a second escape from the water. Impairment in this and other
time later that day, the same place-field map will be forms of spatial learning and memory are observed in
recorded from hippocampal place cells on both ses- various transgenic mouse models of AD.
sions. Testing old rats in this two-session recording Dale Schenk and colleagues (1999) made a re-
procedure, Barnes and her colleagues (1997) ob- markable discovery that vaccinations with A peptide
served a bimodal distribution of responses. For two- can dramatically reduce amyloid deposition in trans-
thirds of the double-session recordings, the old rats genic mouse models of AD. A peptide vaccination
retrieved the same map on both occasions, perform- prevents spatial learning and memory loss (Morgan
ing normally as young rats. However, on one-third of et al., 2000). The long-term behavioral results of A
the double-session recordings, the old rats exhibited peptide vaccinations indicate that the behavioral pro-
a complete rearrangement of the place-field map be- tection of the vaccinations is task-specific, with preser-
tween the two sessions. They apparently retrieved the vation of hippocampal-associated spatial-memory
wrong map on one of the sessions. This failure to re- tasks most likely to occur (Arendash et al., 2001).

Evidence suggests that normal aging affects forty-five-month-old rabbits: Behavioral learning and hippo-
mammalian eyeblink conditioning through age- campal unit activity. Neurobiology of Aging 8, 101108.
related deficits in the cerebellum and hippocampus. Diana S. Woodruff-Pak
Age-related changes in spatial learning and memory
also rely on hippocampal mechanisms. AD exacer-
bates impairment in learning and memory and pro-
foundly disrupts hippocampal function early in its AGING AND MEMORY IN HUMANS
course. Transgenic models of AD provide a means to
test therapeutic interventions, such as vaccination One of the commonest complaints of older people is
with A peptide, that might protect against the cogni- that their memory is not what it used to be. The validi-
tive and neural impairment characteristic of this neu- ty of such subjective reports is borne out by the scien-
rodegenerative disease. tific literature: Memory performance does decline as
a function of the normal aging process in healthy
adults, although the decline is much more evident
See also: AGING AND MEMORY IN HUMANS with some materials and tasks than it is with others.
This variability has given researchers useful clues to
Bibliography the specific memory components or processes that
Arendash, G. W., Gordon, M. N., Diamond, D. M., Austin, L. A., are particularly vulnerable to the effects of aging.
Hatcher, J. M., Jantzen, P., DiCarlo, G., Wilcock, D., and Mor- Some of this work is described below.
gan, D. (2001). Behavioral assessment of Alzheimers trans- Nearly all the studies described use the cross-
genic mice following long-term A vaccination: Task specifici-
ty and correlations between A deposition and spatial sectional method of age comparison; that is, a group
memory. DNA and Cell Biology 20, 737744. of young adults (often college students in their early
Barnes, C. A. (2001). Plasticity in the aging central nervous system. twenties) is compared with a group of older adults
International Review of Neurobiology 45, 339355. (usually community-dwelling volunteers in their six-
Barnes, C. A., Suster, M. S., Shen, J., and McNaughton, B. L. ties and seventies). Additionally, some studies incor-
(1997). Cognitive map multistability in aged rat hippocam-
pus. Nature 388, 272275.
porate middle-aged groups of people in their forties
Chen, L., Bao, S., Lockard, J. M., Kim, J. J., and Thompson, R. F. and fifties. The cross-sectional method is much more
(1996). Impaired classical eyeblink conditioning in cerebellar practicable than within-subject longitudinal studies,
lesioned and Purkinje cell degeneration (pcd) mutant mice. but it does leave open the possibility that the differ-
Journal of Neuroscience 16, 2,8292,838. ences observed between the groups may be attribut-
Disterhoft, J. F., and McEchron, M. D. (2000). Cellular alterations
in hippocampus during acquisition and consolidation of hip-
able to causes other than agingto differences in ed-
pocampus-dependent trace eyeblink conditioning. In D. S. ucation or motivation, for example. Obviously
Woodruff-Pak and J. E. Steinmetz, eds., Eyeblink classical condi- researchers take pains to minimize the possibility of
tioning, Vol. 2: Animal models, pp. 313334. Boston: Kluwer Ac- such artifacts, and they do this by matching the
ademic Publishers. groups by educational level, by vocabulary (a rough
McEchron, M. D., Weible, A. P., and Disterhoft, J. F. (2001). Aging
and learning-specific changes in single-neuron activity in CA1
measure of verbal intelligence), and by other indica-
hippocampus during rabbit trace eyeblink conditioning. Jour- tors of intelligence and socioeconomic status. Many
nal of Neurophysiology 86, 1,8391,857. crucial experimental results take the form of interac-
Morgan, D., Diamond, D. M., Gottschall, P. E., Ugen, K. E., Dick- tions between age and some experimental variable;
ey, C., Hardy, J., Duff, K., Jantzen, P., DiCarlo, G., Wilcock, that is, one condition of the experiment is associated
D., Connor, K., Hatcher, J., Hope, C., Gordon, M., and Aren-
dash, G. W. (2000). A peptide vaccination prevents memory with large age-related differences, whereas another
loss in an animal model of Alzheimers disease. Nature 408, condition is associated with much smaller differences,
982985. even though the same subjects are used. The finding
Schenk, D., Barbour, R., Dunn, W., Gordon, G., Grajeda, H., of such differential effects makes it harder to argue
Guido, T., Hu, K., Huang, J., Johnson-Wood, K., Kahn, K.,
that group differences are a function, say, of reduced
Kholodenko, D., Lee, M., Liao, Z., Lieberburg, I., Motter, R.
Mutter, L., Soriano, F., Shopp, G., Vasquez, N., Vandevert, motivation in the older sample.
C., Walker, S., Wogulis, M., Yednock, T., Games, D., and Seu- Like other experimental explorations of individ-
bert, P. (1999). Immunization with amyloid- attenuates Al-
ual differences in memory ability, the work on aging
zheimer-disease-like pathology in the PDAPP mouse. Nature
400, 173177. has been carried out within various theoretical frame-
Thompson, R. F. (1988). The Rosetta stone for brain substrates of works and with a view to establishing some theoretical
age-related deficits in learning and memory? Neurobiology of point. In this brief article it is not possible to go into
Aging 9, 547548. details of theoretical motivation in most cases, but, in
Woodruff-Pak, D. S., Cronholm, J. F., and Sheffield, J. B. (1990).
summary, four main age-related changes that argu-
Purkinje cell number related to rate of eyeblink classical con-
ditioning. NeuroReport 1, 165168. ably underlie changes in memory performance are a
Woodruff-Pak, D. S., Lavond, D. G., Logan, C. G., and Thompson, decline in general processing speed; a decline in pro-
R. F. (1987). Classical conditioning in three-, thirty-, and cessing resources or attentional energy (Craik and

Byrd, 1982); an age-related reduction in the efficien- of time and place are crucial. There are different
cy of inhibitory processes; and an impairment in the views on whether episodic and semantic memory are
executive control of cognitive processing. These four different memory systems or whether they simply
theoretical viewpoints, along with supporting evi- exemplify different degrees of abstraction from the
dence, are discussed in Kester, Benjamin, Castel, and original events that gave rise to the encoded knowl-
Craik (2002); and in Zacks, Hasher, and Li (2000). edge (Craik, 2002). Whatever the resolution of this
The present article focuses on empirical work since debate, older adults typically show fewer losses in se-
the 1980s, pointing out the implications for theory mantic memory than in episodic memory, provided
where relevant. that the knowledge in question is used on a regular
basis. Thus, older people show minimal losses in vo-
cabulary and in knowledge of facts and concepts.
Indirect Memory Tests On the other hand, some aspects of semantic
When we think about memory, it is usually in the memory do appear to decline with agethe ability to
sense of the conscious retrieval of a past event or the learn completely new facts, for example. Even well-
retrieval of a previously learned fact. However, there learned material can be more difficult to retrieve for
are also many cases in which previous events can af- older adults, and difficulty in remembering names is
fect present behavior in the absence of conscious possibly the most frequent age-related memory com-
awareness of the past event in question. Such cases of plaint. It is unclear, however, whether names show a
implicit memory are revealed by indirect memory disproportional impairment with age (Maylor, 1997);
testsexamples include word-stem completion and a more general statement might be that older adults
word-fragment completion. In word-stem comple- have sporadic difficulty in retrieving any information
tion, the participant is given the first few letters of a that they use infrequently. Having a word or name
word (e.g., MAR___ or DRA___) and asked to com- on the tip of the tongue is an experience that occurs
plete the stem with the first word that comes to mind. more frequently as people grow older, and older
Word-fragment completion is similar; here the partic- adults report less partial information about the target
ipant is given words with letters missing and is asked word. In addition, speed of retrieval slows with ad-
to complete them (e.g., M _ _ K _ T or _ R A _ _ R). vancing age, in line with the slowing of many other
The general finding in such experiments is that par- cognitive processes. In summary, then, whereas the
ticipants can complete the words more readily if they representation of learned knowledge remains reason-
have studied the target words (e.g., MARKET, ably intact into old age, older adults experience occa-
DRAWER) previously, even though the person may sional difficulty in assessing that knowledge.
be quite unaware that the words are drawn from the
studied list. This phenomenon is known as prim-
ing, and it has typically been found that age differ- Episodic Memory
ences in tests of priming are much smaller than those Episodic memory refers to the ability to recollect
found in direct tests (Fleischman and Gabrieli, 1998). specific events, and this form of memory has been ex-
A similar discrepancy between direct and indirect tensively studied in the laboratory using tests of recall
tests is found in amnesic patients, who do poorly on and recognition. Many such studies have used rather
explicit or direct tests but comparatively well on im- artificial materialslists of unrelated words, for ex-
plicit or indirect tests. These results are sometimes ample. The benefit has been greater experimental
taken as evidence that encoding processes are there- control over encoding and retrieval processes, but a
fore intact in elderly individuals and amnesic pa- possible drawback is that the principles emerging
tients, and that the observed decrements are failures from such studies might not apply to real-life remem-
of retrieval. But encoding processes may be somewhat bering. Work carried out beginning in the 1980s on
impaired in older people and amnesics; they might be autobiographical memory has allayed these fears to a
sufficient to support later indirect tests, but insuffi- large extent, however. In these studies, memory for
cient to support later direct tests. personally experienced events in the subjects life
have shown that real-life memories are affected by the
same factors and subject to the same laws as are mate-
Semantic Memory rials learned in the laboratory. Age-related differ-
Semantic memory refers to a persons learned ences in autobiographical memory have been studied
knowledge of facts and concepts. Typically we are un- by presenting people of different ages with cue words
aware of where and when we first learned that Paris such a flag or school and asking participants to gener-
is the capital of France, the meaning of rhinoceros, and ate a personal memory that each word evokes. For
that 7 x 9 = 63; the notion of semantic memory thus adults of all ages, recent memories were generated
stands in contrast to episodic memory, where details most often, and the incidence of recollected events

declined from the present to the past. One interesting gion is associated with the processing of semantic in-
exception to this general trend is that all adults formation and with good later memory of processed
showed a disproportionate bump of generated items. Other studies have shown that, whereas young
memories from late teenage and early adult years adults activate regions that are predominantly lateral-
(Rubin, Wetzler, and Nebes, 1986). Presumably this ized in the right prefrontal cortex during retrieval,
bump reflects the many important career-related and older adults show activations in both right and left
emotionally significant events that occur during this prefrontal regions when retrieving information. The
period of our lives. additional left-sided activations in older people may
reflect the brains attempt to compensate for the de-
Episodic memory for events occurring in the last
clining efficiency of regions that subserve retrieval in
few minutes, hours, or even weeks typically shows
young adulthood. An account of this exciting line of
large age decrements, and much of the effort of cog-
research is provided by Prull, Gabrieli, and Bunge
nitive aging researchers has been directed to under-
standing the factors underlying this problem. One set
of findings points to an age-related decline in the effi- One further topic pertaining to episodic memory
ciency of retrieval processes. Several investigators is the ability to remember the source of encoded infor-
have shown larger age differences in tests of free re- mation, or details of the context in which an event oc-
call (in which participants must recall a list of words curred. Everyone has experienced the situation in
or a paragraph of text with no cues or reminders) which a persons face seems very familiar, and yet we
than in tests of recognition memory (in which partici- cannot say how we know the person. Usually it turns
pants must pick out the originally studied items from out that we have encountered the familiar person in
a mixed set of targets and distractors). Details of these a very different context from his or her habitual loca-
experiments are summarized by Kester, Benjamin, tion, leading George Mandler to refer to the experi-
Castel, and Craik (2002). The fact that older adults ence as the butcher on the bus phenomenon. A sim-
have difficulty recalling studied items but can recog- ilar failure of memory for context occurs when we
nize them suggests that the difficulty lies at the re- know a fact but cannot recollect whether someone
trieval stage, although recognition is usually the easi- told us the information or whether we read it in the
er test. It seems likely that there are also problems at newspaper or heard it on the radio. Older adults are
encoding, however. One piece of evidence supporting particularly prone to such failures to bind contextual
the encoding hypothesis is that, when encoding pro- information to the core aspects of the event, or to the
cesses are guided appropriately by means of ques- item of information (see Kester, Benjamin, Castel,
tions that emphasize the semantic aspects of the item, and Craik, 2002, for details). The effects are also seen
age-related differences are often reduced (Craik and during the output of information; for example, an
older person may be cued by a conversational com-
Jennings, 1992). One similarity between encoding
panions background to exclaim, Ahyou will be in-
and retrieval is thus that, when appropriate processes
terested to hear this! and then proceed to retell the
are supported or guided by additional information
same story that the hapless listener has endured many
(e.g., semantic orienting tasks at encoding, retrieval
times previously.
cues or a recognition test at retrieval), age differences
are typically reduced. This pattern of findings led What lies behind the phenomenon of source for-
Craik (1983) to suggest that, whereas unsupported getting? Some researchers have linked the failure to
encoding and retrieval processes are often inefficient bind contextual details to core features to an ineffi-
in older people, possibly due to a reduction in avail- ciency of frontal lobe function. Other cognitive
able processing resources, these inefficiencies can be neuroscientists have suggested that the hippocampus
overcome through the environmental support pro- is centrally concerned with such binding functions
vided by the experimental situation or by the context (Prull, Gabrieli, and Bunge, 2000). At the behavioral
of a persons familiar surroundings. level, source forgetting may be viewed as one instance
of a general age-related difficulty of association or in-
The conclusion that age-related difficulties in ep- tegration. Naveh-Benjamin (2000) lays out convinc-
isodic memory are consequences of impaired process- ing evidence that older adults have particular difficul-
ing at both encoding and retrieval is supported by ty remembering the associative links between items of
findings from studies of functional neuroimaging. information, although the items themselves may be
Studies using PET (positron emission tomography) remembered quite well.
and fMRI (functional magnetic resonance imagery)
have shown that older adults exhibit less activation of
the ventral left prefrontal cortex during memory en- Short-Term and Working Memory
coding than do their younger counterparts. Many The phrase short-term memory has unfortunately
studies have demonstrated that this left prefrontal re- been used in a number of slightly different ways, and

this can give rise to confusions about findings. Clini- Prospective Memory
cians typically use the phrase to mean recent memo- Prospective memory refers to the situation in
ryevents that have happened in the last few hours which a person intends to carry out some action at a
or dayswhereas experimental psychologists have future time and then either performs the action suc-
used the phrase to refer to information still held in cessfully or forgets to perform it. Such situations are
mind, as when we look up a telephone number and common in everyday life, as are failures of prospec-
rehearse the information until we have dialed it. This tive memoryforgetting to make a phone call, mail
latter type of memory (sometimes also referred to as a letter, or to pass on a message, for example. Re-
primary memory) shows very little decline with age. searchers have made the useful distinction between
Similarly, memory spanthe longest list of digits or event-based and time-based prospective memory, the
words that a person can repeat back accurately first referring to situations in which the intended ac-
declines only slightly from the twenties to the tion should be cued by an event, such as seeing the let-
eighties. If short-term memory is used in the first sense, ter to be mailed or meeting the colleague for whom
however, it falls into the general category of episodic the message was intended. On the other hand, time-
memory, and substantial age-related decrements are based prospective actions are cued by times: I should
found, as discussed in the previous section. call home at 3:30 P.M., for example.
The term working memory has been adopted wide- Prospective memory failure generally increases
ly to refer to information held and manipulated in with age. As one example, Mntyl and Nilsson
mind. Thus, solving a verbal problem or performing (1997) reported a study in which participants were
mental arithmetic is considered to involve working asked to remind the experimenter at the end of a test-
memory (WM). In this sense WM incorporates execu- ing session that they should sign a form. Successful
tive processes as well as relatively automatic auxiliary performance dropped from 61 percent of partici-
systems such as the articulatory loop and visuo-spatial pants aged thirty-five to forty-five to only 25 percent
sketchpad (Baddeley, 1986). With regard to aging, of participants aged seventy to eighty. Other research
performance relying largely on auxiliary systems demonstrates that older adults do worse on time-
holds up well; memory span falls into this category. based than on event-based, prospective memory
But when good performance requires executive pro- tasks, arguably because the former type of task is less
cesses or complex manipulations of information held well supported by environmental cues (Craik, 1983).
in mind, then older people typically do less well than Finally, older people do better on real-life prospec-
their younger counterparts (Craik and Jennings, tive memory tasks than on laboratory-based tasks
1992; Zacks, Hasher, and Li, 2000). It seems possible (Kester, Benjamin, Castel, and Craik, 2002). This
that this age-related decline reflects the reduced effi- finding may reflect greater motivation on the part of
ciency of frontal lobe processes in older adults older adults, or it may reflect their greater use of daily
(Glisky, Polster, and Routhieaux, 1995). structures and routines.
Older adults often have difficulty dealing with
dual-task situations in which they must divide their at-
tention between two simultaneous activities. In one
such demonstration Anderson, Craik, and Naveh- Memory performance does decline with age, but
Benjamin (1998) found that, when younger and older the decline is greater in some tasks than in others.
adults divided their attention between a memory task Performance is often poor on episodic memory tasks,
and an ongoing reaction-time task, memory perfor- especially if the person must recall the information
mance dropped equally for the younger and older without cues. Performance is also comparatively poor
groups (relative to performing the memory task on its on source memory, on prospective memory, and on
own) but that performance on the reaction-time task working memory tasks. On the other hand, memory
dropped much more for older than for younger for general knowledge and for routine activities holds
adults, especially during the retrieval phase of the up well with age, as does primary memory for infor-
memory task. In a similar demonstration, Lindenber- mation held briefly in mind. Finally, environmental
ger, Marsiske, and Baltes (2000) measured walking support from familiar surroundings can be particu-
accuracy and speed in adults of different ages while larly helpful as an aid to remembering in older adults.
they learned a list of words. They found greater dual-
task costs in the older group, partly reflecting the in- See also: AGING AND MEMORY IN ANIMALS;
creased need for executive processes in word learning AUTOBIOGRAPHICAL MEMORY; EPISODIC
but also partly reflecting the older adults greater MEMORY; INDIVIDUAL DIFFERENCES IN
need to deploy executive processes to control accu- LEARNING AND MEMORY; TIP-OF-THE-TONGUE

Bibliography formance on the basis of its own experience. Also

Anderson, N. D., Craik, F. I. M., and Naveh-Benjamin, M. (1998). called learning procedures, methods, or rules, learn-
The attentional demands of encoding and retrieval in youn- ing algorithms are key components of mathematical
ger and older adults: 1. Evidence from divided attention
models of animal learning and of technological de-
costs. Psychology and Aging 13, 405423.
Baddeley, A. D. (1986). Working memory. London: Oxford Univer- vices engineered to improve their behavior as they
sity Press. operate. Learning algorithms have been studied ex-
Craik, F. I. M. (1983). On the transfer of information from tempo- tensively in artificial intelligence, psychology, compu-
rary to permanent memory. Philosophical Transactions of the tational neuroscience, statistics, and engineering.
Royal Society of London, ser. B, 302, 341359.
They are used in many applications in science, indus-
(2002). Human memory and aging. In L. Bckman and C.
von Hofsten, eds., Psychology at the turn of the millennium, Vol. try, and finance that require fitting models to or find-
1: Cognitive, biological, and health perspectives, pp. 261280. ing patterns in collections of data or that require
Hove, Sussex, UK: Psychology Press. learning skills needed to solve specific tasks. Learning
Craik, F. I. M., and Byrd, M. (1982). Aging and cognitive deficits: algorithms play major roles in artificial neural net-
The role of attentional resources. In F. I. M. Craik and S. Tre-
work systems, where they provide rules for adjusting
hub, eds., Aging and cognitive processes, pp. 191211. New
York: Plenum. the strengths, or weights, of connections between net-
Craik, F. I. M., and Jennings, J. M. (1992). Human memory. In F. work elements. These connections correspond to syn-
I. M. Craik, and T. A. Salthouse, eds., The handbook of aging apses by which neurons communicate with other neu-
and cognition, pp. 51110. Hillsdale, NJ: Erlbaum. rons and with sensory and motor mechanisms.
Fleischman, D. A., and Gabrieli, J. D. E. (1998). Repetition prim-
ing in normal aging and Alzheimers disease: A review of find-
ings and theories. Psychology and Aging 13, 88119.
Glisky, E. L., Polster, M. R., and Routhieaux, B. C. (1995). Double
Hebbian Learning
dissociation between item and source memory. Neuropsy- In 1949 the psychologist Donald Hebb hypothe-
chology 9, 229235. sized that when a neuron, A, repeatedly and persis-
Kester, J. D., Benjamin, A. S., Castel, A. D., and Craik, F. I. M.
tently takes part in firing another neuron, B, the syn-
(2002). Memory in the elderly. In A. Baddeley, B. Wilson, and
M. Kopelman, eds., Handbook of memory disorders, 2nd edition. apses by which A stimulates B are strengthened. It
London: Wiley. then becomes easier for neuron A to fire neuron B.
Lindenberger, U., Marsiske, M., and Baltes, P. B. (2000). Memoriz- Consequently, any event that stimulates neuron A
ing while walking: Increase in dual-task costs from young may also stimulate neuron B and thus become associ-
adulthood to old age. Psychology and Aging 15, 417436.
ated with the consequences of Bs firing as well as with
Mntyl, T., and Nilsson, L. G. (1997). Remembering to remember
in adulthood: A population-based study on aging and pro- other events that stimulate B. This is one of the earli-
spective memory. Aging, Neuropsychology, and Cognition 4, 81 est and most influential ideas contributing to neural-
92. network learning algorithms.
Maylor, E. A. (1997). Proper name retrieval in old age: Converging
evidence against disproportionate impairment. Aging, Turning this hypothesis into a complete learning
Neuropsychology, and Cognition 4, 211266. algorithm requires precise definitions of the variables
Naveh-Benjamin, M. (2000). Adult age differences in memory per- involved and how they interact. One of the simplest
formance: Tests of an associative deficit hypothesis. Journal of ways to do this is to consider the situation in which
Experimental Psychology: Learning, Memory, and Cognition 26,
neuron B receives input from other neurons, labeled
Prull, M. W., Gabrieli, J. D. E., and Bunge, S. A. (2000). Age-related 1, 2, . . . , n, any of which can play the role of neuron
changes in memory: A cognitive neuroscience perspective. In A in Hebbs hypothesis. Let XB(t) be a positive num-
F. I. M. Craik and T. A. Salthouse, eds., The handbook of aging ber representing the activity level (the instantaneous
and cognition, 2nd edition, pp. 91153. Mahwah, NJ: Erlbaum. rate of firing) of neuron B at time t. Similarly, for i =
Rubin, D. C., Wetzler, S. E., and Nebes, R. D. (1986). Autobio-
1, 2, . . . , n, let Xi(t) represent the activity levels at
graphical memory across the adult life-span. In D. C. Rubin,
ed., Autobiographical memory, pp. 202221. Cambridge, UK: time t of neurons 1, 2, . . . , n (see Figure l). In repre-
Cambridge University Press. senting how the activity of neuron B depends on the
Zacks, R. T., Hasher, L., and Li, K. Z. H. (2000). Human memory. activities of the other neurons, the simplest assump-
In F. I. M. Craik and T. A. Salthouse eds., The handbook of tion is that XB(t) is a weighted sum of the activities of
aging and cognition, pp. 293357. Mahwah, NJ: Erlbaum.
the other neurons, where each weight represents the
Fergus I. M. Craik current strength of a synapse. If Wi(t) is the strength
at time t of the synapse by which neuron i influences
neuron B, this means that

XB(t) = W1(t)X1(t) + . . . + Wn(t)Xn(t). (1)

Learning algorithms are sets of rules, usually ex- Hebbs hypothesis suggests how the synaptic
pressed using mathematical equations or computer strengths change over time, depending on the other
instructions, that enable a system to improve its per- variables. The extent to which neuron i takes part in

firing neuron B at a time t is often represented by the Figure 1

product of the activity levels of neuron B and neuron
i: XB(t) Xi(t). This product is large when the activity
levels of both neurons are high, and it is small when
the activity level of one or both neurons is close to
zero. Thus, if neuron i persistently takes part in firing
neuron B, this product will be large for many times
t. This leads to a rule for changing synaptic strengths
that can be written for each synapse i as:

Wi(t + t) = Wi(t) + cXB(t)Xi(t), (2)

where t is a small time increment and c is a small pos-

itive number. Selecting values for t and c determines
how rapidly the synaptic strengths change. According
to Equation 2, at any time, t, when the activity levels
of neurons i and B are both greater than zero, XB (t)Xi
(t) is greater than zero, and the synaptic strength Wi Neurons 1, 2, . . . n provide input to neuron B.
(t) increases from time t to time t + t. According to
Equation 1, this larger synaptic weight means that
neuron is activity will contribute more to the activity tative and categorical outcomes). Solving it requires
of neuron B in the future. Equations 1 and 2 consti- finding a rule (mathematically, a function) that best
tute a learning algorithm; both are needed to com- fits the data in the training set (where best is pre-
pute the changes in the synaptic strengths when the cisely defined in some way). The resulting rule is then
values Xi (t), i = l, 2, . . . , n, are given for each time used to provide the desired predictions. Many meth-
t. This is the simplest of many learning algorithms ods for solving these problems are formulated as
based on Hebbs hypothesis, and, despite many short- learning algorithms. The training set represents the
comings, it has played an important role in theories experience of the learning system, and the rule being
of learning in neural networks. Brown et al. (1990) learned improves in its ability to provide valid predic-
discuss this and many other Hebb-inspired learning tions as the data is processed. Learning researchers
algorithms and how well they model synaptic proper- think of the training set as a set of examples provided
ties actually observed in regions of the brain such as by a teacher, and they call this process learning by
the hippocampus. example, learning with a teacher, or, most often, su-
pervised learning. Many learning algorithms have
been devised for supervised learning problems, and
Statistical Learning
modern research on the theoretical properties of
Other learning algorithms are motivated by a de- these algorithms is strongly tied to the field of statis-
sire to solve problems that occur frequently in a vari- tics. Hastie et al. (2001) is a good reference for the sta-
ety of fields. One type of problem that has received tistical view of supervised learning algorithms.
much attention, especially in the field of statistics, is
that of function fitting. Suppose you observe the in-
puts and outputs of some system under study and as- Least Mean Square Algorithm
semble a training set of observations (Xi, Zi), i = 1, 2, The least mean square (LMS) algorithm is an in-
. . . , m, where Zi is the systems output for input Xi. fluential supervised learning algorithm proposed by
(The superscripts are used to avoid confusion with the the electrical engineers Bernard Widrow and Marcian
neuron-activity levels used above.) On the basis of Hoff in 1960. Like the Hebbian algorithm described
these data, you would like to predict what the systems above, it can be expressed as a rule for changing the
output will be for new inputs. For example, the system synaptic weights of a neuronlike element, but it re-
might be a medical treatment, with inputs giving fea- quires another variable to provide training input to
tures of patients (such as the results of clinical tests) the element. This signal tells the element what its ac-
and outputs describing treatment outcome (which can tivity should be; it is said to provide the desired out-
be a quantitative measure or a nonnumeric, or cate- puts or target outputs. During learning, input signals
gorical, description, such as recovery /no-recovery). and target outputs are selected from a training set of
You would like to predict the outcomes of treating example input/output pairs. If Z(t) denotes the target
new patients. output at time t for an element whose actual output
In statistics, this is known as a regression or a clas- is X(t), then Z(t) X(t) gives the error in the elements
sification problem (for the respective cases of quanti- output: the discrepancy between what the elements

activity should be and what it actually is. The LMS al- will eventually find a local minimum, meaning a point
gorithm is an error-correction algorithm because it that is better than all points in the immediate neigh-
changes synaptic weights to reduce the sizes of the er- borhood, although a better point may exist. When the
rors gradually over time. elements output is a linear function of its weights, as
given by Equation 1, all local minima are also global
If we suppose that X(t) is a weighted sum of inputs
minima, meaning that the error surface does not get
as given by Equation 1, then the LMS algorithm says
any lower than at these points. This means that the
that a synaptic weight i changes as follows:
LMS algorithm finds a set of weights that produces
the least mean-square error over the relevant inputs,
Wi(t + t) = Wi(t) + c [Z(t) X(t)]Xi(t). (3) thus justifying the algorithms name.

This means that when the elements output, X(t),

is too low, the error is positive and the synaptic Error Backpropagation
weight, Wi(t), increases (assuming both c and Xi(t) are A generalization of the LMS algorithm known as
positive). On the other hand, when the elements out- the error back-propagation algorithm has been influ-
put is too high, the error is negative, and the weight ential in artificial neural network research. (Haykin
decreases. In either case, this change in synaptic [1999] provides details about this and other network
weight tends to make the elements output more learning algorithms.) In its simplest form, this is a
nearly equal to the target output when input element gradient-descent algorithm that applies to networks
i is active in the future. of neuronlike elements that are connected in multiple
The LMS algorithm is closely related to several layers and that are capable of computing nonlinear
other error-correction algorithms. The perceptron functions of the networks input. An error for each el-
learning algorithm differs from the LMS algorithm ement is computed by propagating the error of the
only in using an error that is sensitive to the differ- network as a whole back through the network in the
ence in the signs of the target and actual outputs but direction opposite to the flow of element activity.
is not sensitive to the difference in their sizes. The Mathematically, this process computes the gradient
LMS algorithm is also nearly identical in mathemati- of the error surface at the point corresponding to all
cal form to an influential model of Pavlovian, or clas- the current weights in the network. Unlike the linear
sical, conditioning proposed by psychologists Robert LMS algorithm, the error back-propagation algo-
Rescorla and Allan Wagner (1972). Instead of rithm is not guaranteed to find a best set of weights
changes in synaptic strengths, this model defines because the error surface can have a complex topog-
changes in behavioral variables called associative raphy with many local minima.
strengths that link conditioned stimuli with an uncon-
ditioned response. The LMS algorithm is the basis of
Other Learning Algorithms
many models of the cerebellum, where climbing fi-
bers may provide the error signals. A learning algorithm is unsupervised, or self-
organizing, if its experience comes from a training set
containing inputs but no target outputs. Instead of
Gradient Descent trying to match target outputs, it tries to recode the
Theories of the LMS algorithm and many other inputs according to some built-in principle. For ex-
learning algorithms depend on viewing the learning ample, some unsupervised learning algorithms re-
process as a search for a set of weights that is best ac- code input signals in order to reduce their complexity
cording to some measure of the algorithms perfor- while trying to preserve their information content.
mance. For the LMS algorithm, this measure assigns Often this takes the form of learning how to form
to each possible set of weights the average, or mean, clusters of input signals so that signals within a cluster
of the squares of the errors that would result from are more similar to one another than they are to sig-
using that set of weights for all relevant inputs. If we nals in other clusters. The Hebbian learning algo-
visualize this measure as a surface over weight rithm is most often used for unsupervised clustering
space, the error measure for the current weights cor- in networks where elements compete for the opportu-
responds to a point on this surface, and the LMS algo- nity to represent different inputs.
rithm changes the weights by moving this point down Reinforcement learning algorithms rely on train-
the slope of the surface, thus improving the systems ing input called reward signals, which evaluate the
ability to match the target outputs. This is called a quality of the learning systems behavior without ex-
gradient-descent algorithm because the direction of plicitly telling the system what its outputs should be.
steepest slope is given mathematically by the gradient Reinforcement learning algorithms have to discover
of the error measure. A gradient-descent algorithm how to improve their behavior by trying various out-
ALZHEIMERS DISEASE: Behavioral Aspects 17

puts and comparing the resulting evaluations. This is ALZHEIMERS DISEASE

a form of trial-and-error learning, which should not
be confused with error-correction learning, such as [Alzheimers disease was first described by the German physi-
that performed by the LMS algorithm, which requires cian Alois Alzheimer in 1906. The disease is associated with
target outputs. Many reinforcement learning systems a progressive decline in general cognitive function and a
use temporal difference (TD) algorithms to learn pre- specific impairment in the ability to form new memories (an-
dictions of the amount of reward a reinforcement terograde amnesia). Approximately 4 million individuals in
learning algorithm will accumulate over an extended the United States suffer from Alzheimers disease at an esti-
period. TD algorithms reduce errors in predictions mated cost for treatment and care that exceeds $100 billion
made at different times by adjusting earlier predic- per year. Two entries on Alzheimers disease follow. BEHAV-
IORAL ASPECTS describes the cognitive defects associated
tions to be closer to laterand therefore more reli-
ablepredictions. The computer scientists Richard with the disease and the ways in which it is diagnosed;
Sutton and Andrew Barto (1998) extensively discuss HUMAN DISEASE AND THE GENETICALLY ENGINEERED
reinforcement learning and TD algorithms. Physio- ANIMAL MODELS describes recent progress in elucidating
the genes and proteins implicated in the disease, the develop-
logical experiments by the neuroscientist Wolfram
ment of an animal model for the disease, and some exciting
Schultz (1998) reveal intriguing parallels between the
alternative treatment possibilities. For a discussion of treat-
behavior of TD algorithms and the activity of domp-
ments for Alzheimers disease see PHARMACOLOGICAL
amine-producing neurons in the brain.

Learning algorithms have been devised for very BEHAVIORAL ASPECTS
different purposes: to model hypotheses about neural Alzheimers disease (AD) is the most common cause
mechanisms of learning, to model the learning be- of dementia in the elderly, affecting 50 percent to 70
havior of animals, and to solve important statistical percent of dementia patients. Dementia involves im-
and engineering problems. Despite these different pairments in daily functioning related to a progres-
purposes, the resulting algorithms show a remarkable sive decline in two or more areas of mental ability.
degree of similarity. Perhaps this is not so surprising These mental and functional deficits are traceable to
because animal behavior and its neural basis are na- neuritic plaques and neurofibrillary tangles, which
tures solutions to problems that have much in com- cause cell loss in the brain. As the disease progresses,
mon with those studied by statisticians and engineers. it affects ever larger areas of the brain. Eventually, so
many parts of the brain are involved that patients can-
not move around normally and feed themselves. They
Bibliography then become susceptible to other potentially fatal dis-
Brown, T. H., Kairiss, E. W., and Keenan, C. L. (1990). Hebbian
eases such as pneumonia. The disease can last from
synapses: Biophysical mechanisms and algorithms. Annual Re-
view of Neuroscience 13, 475511.
five to fifteen years. Many drugs are being developed
Hastie, T., Tibshirani, R., and Friedman, J. (2001). The elements of to treat it, but there is no way of preventing the pro-
statistical learning. New York: Springer-Verlag. gression of the disease.
Haykin, S. (1999). Neural networks: A comprehensive foundation.
Upper Saddle River, NJ: Prentice Hall. When the disease was first described by Alois Alz-
Hebb, D. O. (1949). The organization of behavior. New York: Wiley. heimer in 1906, it seemed to affect only persons
Rescorla, R. A., and Wagner. A. R. (1972). A theory of Pavlovian under the age of sixty-five. In the 1960s, researchers
conditioning: Variations in the effectiveness of reinforcement discovered during autopsies that patients whose se-
and non-reinforcement. In A. H. Black and W. F. Prokasy, vere cognitive decline had been attributed to diseases
eds., Classical conditioning, Vol. 2: Current research and theory,
in the blood vessels of the brain (cerebrovascular dis-
pp. 6499. New York: Appleton.
Schultz, W. (1998). Predictive reward signals of dopamine neurons.
ease or hardening of the arteries) in fact exhibited
Journal of Neurophysiology 80, 127. the pathological hallmarks of AD (neurofibrillary tan-
Sutton, R. S. and Barto, A. G. (1998). Reinforcement learning: An in- gles and neuritic plaques). The realization that AD af-
troduction. Cambridge, MA: MIT Press. fects persons of all ages is, therefore, a recent one. Be-
Widrow, B., and Hoff, M. E. (1960). Adaptive switching circuits. In cause of the rarity of the disease among those under
1960 IRE WESCON convention record, pp. 96104. New York:
sixtyless than 2 percentit used to be considered
IRE. Reprinted in J. A. Anderson and E. Rosenfeld, eds.,
Neurocomputing: Foundations of research. Cambridge, MA: MIT
an uncommon disorder. Now it is understood that the
Press, 1988. likelihood of contracting AD increases with age.
Among people sixty-five to eighty-five, its prevalence
Andrew G. Barto is 5 percent to l5 percent; 30 to 47 percent of those
18 ALZHEIMERS DISEASE: Behavioral Aspects

over eighty-five have AD. It is the fourth-leading the concentration of a number of neurotransmitters,
cause of death among those sixty-five or older. the chemicals that are responsible for the transmis-
The earliest symptom of AD in most patients is se- sion of nerve signals in the brain. One neurotransmit-
vere difficulty in learning new information (i.e., an- ter, acetylcholine, which is important for normal
terograde memory impairment), especially increas- human memory, declines by up to 70 percent in cases
ing forgetfulness about day-to-day events. First, of severe impairment. The combination of these
patients may forget a recent event from one week to structural and chemical changes seems to be responsi-
the next, then from one day to the next, and finally ble for the unique severity of anterograde memory
from one minute to the next. AD impairs nearly all as- impairment in AD.
pects of new learning. It is not, for example, limited The cognitive impairments of AD are not con-
to information that the patient is trying to learn (i.e., fined to memory. These other difficulties interact
explicit or episodic memory), as is true in some amne- with the memory impairment to increase its severity.
sic disorders. Patients also have difficulty with implicit Early on in the disease the other main area of impair-
memorylearning information that impinges on ment is the executive functions, which pertain to con-
consciousness in the absence of any effort at master- cept formation, directed attention, and the concur-
ing it. rent manipulation of information. Studies suggest
AD does not uniformly impair all implicit memo- that the early stages of AD present problems with the
ry tasks. The same is true for the ability to learn new concurrent manipulation of information and a conse-
general skills (i.e., procedural knowledge). Patients quent difficulty with tasks that require simultaneous
have difficulty on some kinds of skill learning tasks manipulation of several different components, even
but not others; for example, learning of motor skills when the individual components of the tasks are
is frequently spared. Early in the disease patients deeply ingrained (e.g., preparing meals, paying bills,
memory for remote events is usually intact. As the dis- balancing a checkbook, and so on). That is why it
ease progresses, however, remote memories also re- sometimes seems as if the patients have difficulty with
cede, but in reverse chronological order; the most re- remote memory early in the disease. However, when
cent memories evaporate first. Only the most severe the tasks are broken down into their constituent parts,
impairment obliterates memories of earliest child- patients can perform them.
hood. Recent evidence suggests that problems with ex-
The rapidity of anterograde memory impairment ecutive function are at least in part the result of pa-
in AD is dramatic: It occurs after a delay of ten min- thology in the central portion of the cingulate gyrus.
utes or less after exposure to new information. The Studies in animals demonstrate that damage to the
best way to reveal this rapid loss of information is to middle portion of the cingulate alters executive func-
give a patient something new to learnfor example, tion. Another possible cause of executive-function im-
a storyand to ask him/her to state how much is re- pairment may be the loss of the corticocortical fibers,
membered immediately, and then after a delay. The which connect different parts of the cortex to one an-
delay should be less than ten minutes, because after other. The partial degeneration of this intracortical
ten minutes most types of patients, and even normal projection system in the early stages of AD could im-
older persons, forget a substantial amount of infor- pair the performance that requires the rapid and si-
mation. For example, patients with other forms of de- multaneous integration of multiple types of informa-
mentia, such as Picks disease or Huntingtons dis- tion.
ease, also have difficulty learning and retaining new The further progression of AD often severs the
information, but if delays last less than ten minutes, meaningful associations among wordshence a loss
their retention is significantly better than that of AD of semantic knowledge or semantic memory. Patients
patients. If delays are longer than ten minutes, then with AD begin to have difficulty with a number of lin-
normal patients and those with other forms of de- guistic tasks when they are mildly-to-moderately im-
mentia do not differ significantly from one another. paired. These difficulties affect the ability to produce
Two types of changes in the brain are primarily the name of an object when shown the object itself or
responsible for the severe anterograde memory im- a picture of the object (confrontation naming). They
pairment of AD. The first is damage to the brain re- also affect the ability to produce rapidly a series of
gions in the medial temporal lobe that are essential words that belong to a particular category, such as
for normal memory, specifically the entorhinal cortex vegetables or words starting with the letter s (verbal
and the hippocampus. Even in very mildly impaired fluency).
patients, there is approximately a 30-percent neuro- Some researchers argue that AD involves a break-
nal loss in the entorhinal cortex, the primary input down in the structure of semantic knowledge and that
path to the hippocampus. There are also declines in patients actually lose knowledge they once possessed.
ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models 19

Others argue that semantic knowledge remains intact McKhann, G., Drachman, D., Folstein, M. F., Katzman, R., Price,
in mildly and moderately impaired patients but is D., and Stadlan, E. (1984). Clinical diagnosis of Alzheimers
disease. Report of the NINCDS-ADRDA Workgroup under
more difficult to access; thus, when one asks them to the auspices of Department of Health and Human Services
intentionally search their semantic memory for infor- Task Force. Neurology 34, 939944.
mation, they have difficulty, but if their semantic Petersen, R., Smith, G., Waring, S., Ivnik, R., Tangalos, E., and
knowledge is evaluated indirectly, it appears to be Kokmen, E. (1999). Mild cognitive impairment, clinical char-
preserved. It has been difficult to resolve this debate acterization and outcome. Archives of Neurology 56, 303308.
because research, although intensive, often yields Marilyn S. Albert
contradictory findings.
AD is still difficult to diagnose during life. No test
can definitively verify the onset of the disease short of HUMAN DISEASE AND THE GENETICALLY
examining brain tissue through a biopsy. Researchers ENGINEERED ANIMAL MODELS
have therefore developed a number of conventions to
Alzheimers disease (AD) represents a great challenge
communicate the degree to which patients have been
for science and medicine because of its prevalence,
examined and the diagnostic criteria that the patients
cost, lack of reliable treatments, and often devastating
meet. Patients who have been carefully examined and
impact on individuals and caregivers. This age-
who meet clinical research criteria but have not had
associated chronic illness involves genetic risk factors,
a brain biopsy are said to have probable AD, or proba-
a well-defined clinical syndrome with a progressive
ble dementia of the Alzheimer type. Patients who
course, evidence of dysfunction and/or death of popu-
have had dementia during life and in whom results of
lations of neurons, pathological and biochemical ab-
an examination of brain tissue meet pathological re-
normalities, and intra- or extra-cellular protein ag-
search criteria for the disease are said to have definite
gregates (Price, Tanzi, Borchelt, and Sisodia, 1998).
Patients become severely disabled and often die of in-
Because novel treatments are under investiga- tercurrent illnesses. There are treatments for symp-
tion, there has been increasing interest in determin- toms but no cure. However, recent research, particu-
ing the earliest possible diagnosis of AD. The term larly in animal models, has begun to provide new
mild cognitive impairment (MCI) has been developed to insights into the mechanisms of Alzheimers disease
describe individuals with evidence of a functional dif- and has identified new targets for therapy.
ficulty in daily life that does not warrant a diagnosis
of dementia. A large number of individuals with MCI
(approximately 15 percent per year) are later diag- Clinical-Pathological Features of
nosed with AD. Drug trials are underway to see if Alzheimers Disease
treating patients with MCI reduces the likelihood that In most cases of AD, the initial impairments of
they will be diagnosed with AD. memory and cognition appear gradually during the
AD is a common and serious disorder that is seventh decade. The accuracy of clinical diagnoses
caused by damage to the structure and function of the improved from 1980 to 2000, and early diagnosis will
brain. There are several hypotheses concerning the become increasingly important as mechanism-based
underlying cause of AD, but these remain unproved, treatments become available.
and there are no effective treatments. The disease AD involves the brain (and not other organs) and
produces a gradual decline in cognitive function. The certain neuronal populations are selectively vulnera-
most common early symptom is a dramatic loss of new ble. AD results from the selective degeneration of
information after a brief delay, but AD can impair all nerve cells in the brains regions and neural circuits
aspects of mental ability. that are critical for memory, cognitive performance,
and personality (Albert, 1996). The dysfunction and/
TREATMENTS OF MEMORY DEFICITS or death of these neurons reduces the numbers of ge-
neric and transmitter specific-synaptic markers in
Bibliography their target fields; the disruption of synaptic commu-
Albert, M., and Moss, B. (1999). Early features of Alzheimers dis- nication in affected regions/circuits can lead to men-
ease. In A. Peters and J. Morrison, eds. Cerebral Cortex, Vol. 14, tal impairments and, finally, severe dementia.
pp. 461471, New York: Plenum.
Gomez-Isla, T., Price, J., McKeel, D., Morris, J., Growdon, J., and AD typically involves intracellular or extracellular
Hyman, B. (1996). Profound loss of layer II entorhinal cortex protein aggregates in brain. Neurofibrillary tangles
neurons occurs in very mild Alzheimers disease. Journal of (NFTs), inclusions located within cell bodies and
Neuroscience 16, 4,4914,500.
Katzman, R. and Kawas, C. (1984). The epidemiology of dementia proximal dendrites, are composed of poorly soluble
and Alzheimers disease. In R. D. Terry, R. Katzman, and K. paired helical filaments (PHF), which are, in turn,
L. Bick, eds., Alzheimer disease. New York: Raven. composed principally of hyperphosphorylated isofor-
20 ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models

ms of tau. PHF are also present in dystrophic neu- thought to be either tumor necrosis factor (TNF)
rites, the filamentous swellings of distal axons/ converting enzyme (TACE) or a disintegrin and met-
terminals (usually seen in proximity to A deposits). alloproteinase 10 (ADAM 10), which cut between resi-
Perturbations related to hyperphosphorylated tau dues 16 and 17 of A, and by BACE2, a protease shar-
seem to play a role in disturbances in intracellular ing features with BACE1, but cleaving APP after
transport. residues 19 and 20 of A (i.e., within the A domain).
The extracellular aggregates in brain are abnor- These different endoproteolytic cleavages generate
mal accumulations of A, a 4kD pleated sheet amy- various C-terminal peptides, including the APP intra-
loid peptide, derived by - and -secretase cleavages cellular domain (C60), which may play a role in the
of the amyloid precursor protein (APP). Levels of A activation of transcription.
are elevated in brain, and A monomers form oli- A variety of APP mutations, including APPswe (a
gomers and multimers that assemble into protofila- double mutation at the N-terminus of A) and APP-
ments and then fibrils. Eventually, A fibrils are de- 717 (near the C-terminus of A), have been reported
posited as the amyloid cores of neuritic or senile in cases of FAD. These mutations, strikingly situated
plaques, which are complex structures also containing near several secretase cleavage sites, are proamyloi-
dystrophic neurites, astrocytes, and microglia. dogenic, and cells that express mutant APP show ab-
Plaques are preferentially localized to the cortex, hip- errant APP processing: the APPswe mutation, which
pocampus, and amygdala. enhances BACE1 cleavage, is associated with elevated
The levels and distributions of APP and its cleav- levels of A; the APP 717 mutations, which affect -
age enzymes in neurons lead to the selective appear- secretase activity, lead to a higher secreted fraction of
ance of A in brain. It seems that toxic A peptides, longer, more toxic A peptides (A42) relative to cells
particularly oligomers, accumulate near synapses and that express wild-type APP (Price, Tanzi, Borchelt,
may impair transsynaptic communication, eventuat- and Sisodia, 1998; Citron et al., 1992).
ing in the disruption of synaptic connections between
PS1 and PS2
neurons and their targets (other nerve cells). Nerve
cells are functionally damaged, changes occur in tau Localized to chromosomes 14 (PS1) and 1 (PS2),
phosphorylation; microtubule stability is compro- respectively, these genes encode highly homologous
mised; intracellular transport processes are impaired; 43- to 50-kD proteins with multiple transmembrane
intracellular PHF appear; and cell geometry is al- (TM) domains. Oriented toward cytoplasm are a hy-
tered, with synapses, axon terminals, and dendrites drophilic acidic loop region, an N-terminal region,
appearing to be most vulnerable. The neuron is inca- and a C-terminal domain. PS1 is synthesized as an
pable of performing its normal functions for a signifi- 42- to 43-kD polypeptide, but the preponderant
cant interval before the ensuing demise of the cell. PS1-related species that accumulate in vitro and in
vivo are 27- to 28-kD N-terminal and 16- to 17-kD
C-terminal derivatives, which accumulate and/or as-
Mutant Genes/Proteins Implicated in sociate in a 1:1 stochiometry and are stable, tightly
Familial Alzheimers Disease (FAD) regulated, and saturable. PS genes are widely ex-
In some individuals with early onset AD, the ill- pressed at low abundance in the CNS.
ness may be inherited as an autosomal dominant with PS1 influences APP processing, but it is not clear
mutations in three different genes: the APP; PS1; and whether PS1 itself is an aspartyl protease (i.e., -
PS2 (Price, Tanzi, Borchelt, and Sisodia, 1998). secretase), is a cofactor critical for the activity of -
APP secretase, or plays a role in trafficking of APP to the
proper compartment for -secretase cleavage (De
Encoded by a gene on chromosome 21, APP is ex-
Strooper et al., 1998).
pressed in many cells and tissues but is particularly
abundant in neurons. This type-1 transmembrane The PS1 gene harbors more than fifty different
protein is cleaved endoproteolytically by an enzyme, FAD mutations in more than eighty families, whereas
-site APP-cleaving enzyme 1 (BACE1), and by an ac- only a small number of mutations have been found in
tivity termed -secretase, which, in concert, gener- PS2-linked families. The vast majority of the abnor-
ate the N- and C-termini of the A peptide, respec- malities in PS genes are missense mutations that re-
tively. The levels and distributions of APP and the sult in single amino-acid substitutions. However, re-
activities of proamyloidogenic cleavage enzymes, par- searchers have found a mutation that deletes exon 9
ticularly BACE1, in neurons seem to be lead to the from PS1 in several different FAD families. The vari-
formation of A in brain. The formation of A 1- ous PS mutations appear to influence -secretase ac-
40,42 is precluded by the endoproteolytic cleavage of tivity and increase the generation of the A 42 pep-
APP within the A sequence by -secretase, now tide.
ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models 21

Transgenic Models of A Amyloidogenesis Consistent with this idea are observations that
APLP2-/- mice appear normal but that mice with ei-
Some of the lines of mutant APP mice, although
ther both APP and APLP2 targeted alleles or both
they do not reproduce the full phenotype of AD, rep-
APLP1 and APLP2 null alleles show significant post-
resent excellent models of A amyloidosis and are of
natal lethality.
great value for testing causal effects of mutant genes,
analyses of pathogenic pathways, determination of BACE1-/- Mice
the molecules participating in A amyloidogenesis,
and identification of therapeutic targets. What follows These null mice are viable and healthy, have no
is a review of selected examples of lines of mice ex- obvious phenotype or pathology, and can mate suc-
pressing autosomal dominant FAD-linked mutant cessfully (Cai et al., 2001). In cortical neurons from
transgenes. BACE1 null embryos, there is no cleavage at the +1
and + 11 sites of A (Cai et al., 2001), and the secre-
Mutant APP Mice tion of peptides is abolished even in the presence
Several promoters have been used to drive the ex- transfected mutant APP transgenes. Moreover, A
pression of an APP minigenes that encode the FAD- peptides are not produced in the brains of BACE1
linked APP mutants (swe and 717) in strains of mice. null mice. These results establish that BACE1 is the
The pathology is influenced by the level of transgene neuronal -secretase required to cleave APP to gener-
product and the specific mutation. The hippocampi ate the N-termini of A they further establish that
and cortices of these mice show elevated levels of A, BACE1 is an excellent therapeutic target for drug de-
diffuse A deposits, and plaques consisting of dystro- velopment for AD.
phic neurites displayed around an A core. Astrocytes
and microglia are clustered in and around plaques. PS1 and PS2 Null Mice
NFT are not apparent. In some lines of mice there To examine the roles of PS1 in development, sev-
may be mild loss of neurons. Some mice show abnor- eral groups have produced PS1-/- mice (4). Homozy-
malities of synapses in hippocampal circuits that pre- gous mutant mice fail to survive beyond the early
cede the deposition of A. In some lines, mice may postnatal period and show severe perturbations in the
exhibit learning deficits, problems in object- development of the axial skeleton and ribs (defects in
recognition memory (related to the number of amy- somitogenesis) that resemble a particular Notch1 null
loid deposits in specific regions) and impairments of phenotype. Because PS1 homologues interact with
alternation/spatial-reference and working memory Notch1, a receptor protein involved in critical cell-
(perhaps related to reductions in synaptic densities in fate decisions during development, it is not surprising
the hippocampus). that cells lacking PS1 show reductions in proteolytic
release of the Notch1 intracellular domain (NICD), a
APPswe/PS1 Mutant Transgenic Mice cleavage that is thought to be critical for Notch1 sig-
Transgenic mice that coexpress A246E HuPS1 naling. Both wild type and mutant human PS1 trans-
and Mouse/Human-APPswe have elevated levels of A genes rescue the spectrum of developmental defects
in the brain and develop numerous amyloid deposits in PS1 null mice. These results indicate that the FAD-
in the hippocampus and cortex (Borchelt et al., 1997; linked PS1 variants retain sufficient normal function
Borchelt et al., 1996). The A deposits are associated to allow normal mammalian embryonic development.
with dystrophic neurites that contain APP, PS1, and With regard to the role of PS proteins in A biology,
BACE1 immunoreactivities; thus, the key participants mutations in PS genes increase the formation of
in amyloidosis appear locally at some of the possible A42, and ablation of PS1 reduces the secretion of A.
sites of formation of A. Significantly, cells from PS1-/- mice show reductions
in the levels of -secretase cleavage products and le-
vels of A
Gene-Targeted Mice Particularly Relevant
to AD
APP and APLP2 Null Mice Vulnerability of Neurons in Alzheimers
Homozygous APP-/- mice are viable and fertile, Disease
but they appear to have subtle decreases in locomotor Among the most challenging mysteries of neu-
activity and forelimb grip strength. The absence of rodegenerative diseases is the identification of factors
substantial phenotypes in APP-/- mice may be related that render neurons susceptible in specific diseases
to the functional redundancy provided by homolo- (the principle of selective vulnerability). For example,
gous amyloid precursorlike proteins (APLP1 and APP and SOD1 are abundant in many cells. Why,
APlP2), molecules expressed at high levels with devel- then, do mutations of genes encoding these proteins
opmental and cellular distributions similar to APP. cause neurological diseases? And why are mutations
22 ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models

in APP associated with the development of a demen- dogenic enzyme, wherease BACE2 is an antiamyloi-
tia syndrome and mutations in SOD1 with an MND dogenic protease, and that the relative levels of
phenotype? Recent research has begun to provide ex- BACE1 and BACE2 are major determinants of A
citing new clues concerning the biological basis for amyloidosis. Significantly, -secretase, which may be
vulnerabilities of neurons. AD serves as an illustra- PS1 (or has its activity influenced by PS), is present in
tion. We suggested that the basis for the vulnerabili- a relatively low level in brain and does not form A
ties of the brain to AD are related to the levels and dis- without BACE1 cleavages.
tributions of APP and its cleavage enzymes in neurons
as opposed to other cells (Cai et al., 2001). APP is one In this model (Cai et al., 2001), the secretion of
of the most abundant proteins in neurons, and avail- A peptides would be highest in neurons/brain as
able evidence indicates that neurons are the principal compared to other cell types/organs because neurons
source of A. In nerve cell, APP is transported within express high levels of BACE1 coupled with low ex-
axons by the fast anterograde system. APP processing pression of BACE2. If the ratio of the level of BACE1
can occur at nerve terminals. In the terminal fields of to BACE2 is a critical factor that selectively predis-
the perforant pathway, BACE1 cleavage generates poses the brain to A amyloidosis, AD would likely in-
soluble C-terminally truncated APPs and amyloido- volve the brain selectively as opposed to other organs.
genic C-terminal fragments. Moreover, in mutant
A seemingly contradictory study shows a high level of
APP transgenic mice, APP, BACE1, and PS1, the key
BACE1 mRNA expression in the pancreas. Since APP
proteins in the formation of amyloid, have been seen
is expressed in the pancreas, why do AD and diabetes
in swollen neurites in immediate proximity to A de-
mellitus not occur together? It now appears that some
posits. These observations conform to the idea that
neurons and their processes, including axon termi- of the pancreatic BACE1 mRNAs are alternatively
nals, are one source of the APP that gives rise to A spliced to generate a BACE1 isoform that is incapable
peptide species. of cleaving APP. Taken together with the observations
that the pancreas possesses low levels of BACE1 and
However, the presence of APP in neurons, al- low amounts of BACE1 activity, these results are con-
though necessary, is not sufficient to explain why the sistent with the view that a high ratio of BACE1 to
brain is particularly vulnerable to A amyloidogenesis
BACE2 activity leads to selective vulnerability of neu-
whereas other organs, such as the pancreas, are not
rons and not pancreatic cells to A amyloidosis. To
. The patterns of APP-cleavage enzymes in different
test this hypothesis at the level of specific cell popula-
cell populations are of equal importance. The cellular
tions, it will be important to define the levels and dis-
distributions, relative levels, and sites of APP cleavage
of BACE1, BACE2, and -secretase are principal de- tributions of BACE1 and BACE2 in specific brain re-
terminants of such vulnerability. Although both gions, circuits, and neurons using specific BACE1 and
BACE1 and BACE2 are expressed ubiquitously, 2 antisera and to attempt to correlate these measures
BACE1 mRNA levels are particularly high in the with the regional vulnerabilities to A amyloidosis
brain and pancreas, whereas the levels of BACE2 seen in AD.
mRNA are relatively low in all tissues except the
brain, where it is nearly undetectable. As indicated
above, A is generated by biochemical pathways in- Treatment in Models of A
volving the endoproteolytic cleavages carried out by Amyloidogenesis
BACE1 and by an activity termed -secretase, which Research on model systems relevant to AD illus-
generate the N- and C-termini of the A peptide, re- trates the value of studies of transgenic and gene-
spectively. Most importantly, BACE1 is the principal
targeted mice for experimental treatments. Although
-secretase necessary to cleave APP at the +1 and
they do not model the full phenotype of AD, these
+11 sites of A in neurons (Cai et al., 2001). In con-
mutant mice represent excellent models of A
trast, BACE2 cleaves APP more efficiently at residues
amyloidogenesis and are highly suitable for analyses
+19 and +20 of APP compared to the +1 site of A.
Significantly, levels of A1-19 and A1-20 are undet- of pathogenic pathways, determination of the molec-
ectable in brain. APP can also be cleaved endoproteo- ular participants in amyloidogenesis, and identifica-
lytically before residue +17 within the A sequence by tion of therapeutic targets. Moreover, they are invalu-
-secretases, either TACE or ADAM10. These able for examining the effects of the introduction
three cleavages within the A domain of APP pre- and/or ablation of specific genes, administration of
clude the formation of A 1- 40,42. Because BACE1 pharmacological agents (secretase inhibitors), and A
is the principal -secretase in neurons (Cai et al., vaccination or passive transfer of A antibody. Re-
2002) and BACE2 may serve to limit the secretion of searchers need to assess the efficacy of anti-A thera-
A peptides, we suggest that BACE1 is a proamyloi- pies in transgenic mice exhibiting tau pathology.

Bibliography Amnesia in the Dissociative Disorders

Albert, M. S. Cognitive and neurobiologic markers of early Al-
zheimer disease (1996). Proceedings of the National Academy of
One major category of functional amnesia occurs
Sciences of the United States of America 93, 13,54713,551. within the context of diagnosable psychopathology,
Borchelt, D. R., Ratovitski, T., Van Lare, J., Lee, M. K., Gonzales, especially the dramatic Dissociative Disorders listed
V. B., Jenkins, N. A., Copeland, N. G., Price, D. L., and Si- in the Diagnostic and Statistical Manual (DSM) (4th
sodia, S. S. (1997). Accelerated amyloid deposition in the edition) of the American Psychiatric Association. In
brains of transgenic mice co-expressing mutant presenilin 1
and amyloid precursor proteins. Neuron 19, 939945.
current diagnostic nosology, this category includes a
Borchelt, D. R., Thinakaran, G, Eckman, C. B., Lee, M. K., Daven- wide variety of syndromes whose common core is an
port, F. Ratovitsky, F. Prada, C.-M., Kim, G. Seekins, S., alteration in consciousness affecting memory and
Yager, D., Slunt, H. H., Wang, R., Seeger, M. Levey, A. I., identity.
Gandy, S. E., Copeland, N. G., Jenkins, N. A., Price, D. L.,
Younkin, S. G., and Sisodia, S. S. (1996). Familial Alzheimers In dissociative amnesia (also known as psycho-
disease-linked presenilin 1 variants elevate A142/140 ratio genic amnesia, limited amnesia), the patient suffers
in vitro and in vivo. Neuron 17, 1,0051,1013. a loss of autobiographical memory for specific past
Cai, H., Wang, Y., McCarthy, D., Wen, H., Borchelt, D. R., Price,
D. L., and Wong, P. C. (2001). BACE1 is the major -secretase
experiences. It sometimes occurs in cases of violent
for generation of A peptides by neurons. Nature Neuroscience crime (interestingly, affecting either victims or perpe-
4, 233234. trators), war neurosis, and other types of posttraumat-
Citron, M., Oltersdorf, T., Haass, C., McConlogue, L., Hung, A. ic stress disorder. Unfortunately, many reports of dis-
H., Seubert, P., Vigo-Pelfrey, C., Lieberburg, I., and Selkoe, sociative amnesia are anecdotal and lack independent
D. J. (1992). Mutation of the -amyloid precursor protein in
familial Alzheimers disease increases -protein production.
corroboration of the purported instigating event.
Nature 360, 672674. Moreover, it is not possible to reliably distinguish
De Strooper, B., Saftig, P., Craessaerts, K., Vanderstichele, H., genuine cases of psychogenic amnesia from simulated
Guhde, G., Annaert, W., Von Figura, K., and Van Leuven, F. cases.
(1998). Deficiency of presenilin-1 inhibits the normal cleav-
age of amyloid precursor protein. Nature 391, 387390. In dissociative fugue (psychogenic fugue, func-
Price, D. l., Tanzi, R. E., Borchelt, D. R., and Sisodia, S. S. (1998). tional retrograde amnesia), the amnesia covers the
Alzheimers disease, genetic studies and transgenic models. whole of the individuals past life and his or her per-
Annual Review of Genetics 32, 461493.
sonal identity; there may also be physical relocation
Donald L. Price (which gives the syndrome its name). The condition
Sangram S. Sisodia may go unnoticed until the patients are asked person-
Revised by Philip C. Wong and Donald L. Price al questions that they cannot satisfactorily answer. Re-
covery typically begins with the patients recognition
of loss of identity. Recovery of identity and memory
per se may occur spontaneously or in response to the
AMNESIA appearance of a relative or other salient cue. When
See: ALZHEIMERS DISEASE: BEHAVIORAL ASPECTS; the fugue is resolved, the patient is typically left with
ALZHEIMERS DISEASE: HUMAN DISEASE AND a limited amnesia covering the period of the fugue.
In dissociative identity disorder (multiple-
personality disorder), a single individual appears to
TRANSIENT GLOBAL; HEAD INJURY; manifest two or more distinct identities, each alter-
KNOWLEDGE SYSTEMS AND MATERIAL- nating in control over conscious experience, thought,
SPECIFIC MEMORY DEFICITS; and action. Before World War II, the typical case in-
REHABILITATION OF MEMORY DISORDERS volved only two or three such ego states; more re-
cent cases have tended to present more alter egos,
leading some to speculate that iatrogenic and so-
ciocultural factors may account for much of the
AMNESIA, FUNCTIONAL multiple-personality epidemic of the 1980s. In genu-
ine cases, the personalities are separated by an amne-
Analyses of learning and memory increasingly at-
sic barrier. The dissociation may be symmetrical, in
tempt to take account of clinical and experimental re-
which each ego state is ignorant of the other(s) or,
search on victims of amnesia. Most of this literature
more commonly, asymmetrical, in which case an ego
has focused on pathologies of memory associated with
state may be aware of some of its counterparts but ig-
demonstrable brain lesions or the administration of
norant of others.
centrally acting drugs. The functional amnesias are a
collection of memory disorders instigated by process- In depersonalization the person believes that he
es that do not result in damage or injury to the brain or she has changed in some way or is somehow unreal;
but that do engender a marked increase in forgetting. in derealization the same beliefs are held about ones

surroundings. Because these beliefs are objectively in- Explicit and Implicit Memory
appropriate, these experiences can be construed as While the functional amnesias by definition im-
disorders of memory: the person fails to recognize pair explicit memory, some anecdotal and experi-
some object, self, or situation with which he or she is mental evidence suggests that the amnesia may spare
objectively quite familiar. Episodes of depersonaliza- implicit memory, or the unconscious influence of past
tion and derealization frequently occur in response to events on subsequent experience, thought, or action.
stress and in association with anxiety disorders; they In dissociative identity disorder, for example, both
may also be induced by psychedelic drugs and occur procedural learning and priming effects may transfer
spontaneously in a substantial proportion of the nor- between personalities, so that one alter ego is influ-
mal population. enced by the experiences of another even though the
Although dissociative disorders have been of in- amnesic barrier prevents conscious recollection. The
terest at least since the time of Freud and Janet, they situation is complicated, however, because not all
rarely have been studied with controlled experimen- forms of implicit memory are equally spared. There
tal procedures. For example, little is known about has been no experimental corroboration of clinical
psychogenic amnesia beyond anecdotes. A few cases claims that special procedures such as hypnosis and
of fugue and multiple personality have been studied barbiturate narcosis can promote conscious access to
in the laboratory, but we have no idea how represen- the lost memories.
tative they are. Nevertheless, the available evidence Among nonpathological amnesias, the dissocia-
suggests a pattern of selective memory deficit in some tion between explicit and implicit memory is especial-
respects resembles that observed in organic amnesia. ly well documented in posthypnotic amnesia. Prever-
Thus, psychogenic fugue impairs memory for past ex- bal infants can show long-term retention of new
periences and other aspects of self-knowledge but learning in a manner that suggests implicit memory.
leaves the patients repertoire of impersonal proce- However, sleep-learning procedures do not appear to
dural and semantic knowledge largely intact. Disso- leave any traces, even in implicit memory.
ciative identity disorder displays a similar pattern.

Trauma, Repression, and Dissociation

Nonpathological Amnesias
The lack of reliable evidence of brain damage has
In other forms of functional amnesia, dramatic fostered a tendency to account for functional amne-
forgetting occurs in the ordinary course of everyday sias in purely psychological terms. Since the nine-
living, albeit with no implication of pathology. For ex- teenth century, repression and dissociation have been
ample, people commonly fail to remember their the favored explanations. Repression, as defined by
dreams and other events of the nights sleep; attempts Freud, is the motivated forgetting of material (typi-
to demonstrate sleep learning have been almost uni- cally, relating to sexual or aggressive ideas and im-
formly unsuccessful. Theoretical accounts of this pulses) that conflicts with physical reality or social
memory deficit usually revolve around encoding fac- sanctions. Dissociation, as discussed by Janet and
tors. For example, one hypothesis holds that sleep in- Prince, is a more adventitious splitting off from
hibits the higher cortical centers that support percep- awareness of a set of percepts, memories, thoughts, or
tual processing. feelings. While Freud argued that repressed contents
Another example of nonpathological functional could be known only by inference (because they were
amnesia is the general paucity of memory for infancy expressed only symbolically, as in dream contents),
and childhood. As with sleep, most theoretical ac- Janet argued that dissociated contents could be recov-
counts of this developmental amnesia focus on encod- ered directly, by hypnosis and other means. Given the
ing factors. For example, infantile amnesia (covering pervasive influence of Freudian psychoanalysis in
the first two years of life) may reflect the childs rela- twentieth-century discourse, the repression thesis
tive inability to encode symbolic and especially lin- long held sway. There has been a subsequent revival
guistic representations of events; even older children of the concept of dissociation, as indicated by adop-
may lack the information-processing capacity to en- tion of Dissociative Disorder as a category in the
code retrievable memories. DSM. An eclectic combination of Freudian and Jun-
gian theories formed the basis of clinical theories
A dramatic form of forgetting known as post- about trauma and memory that emerged in the late
hypnotic amnesia occurs in some hypnotized subjects. twentieth century. These theories, in turn, promoted
In some respects, posthypnotic amnesia may serve as the 1980s revival of therapeutic strategies based on
a laboratory analogue of the dissociative amnesias the recovery and working through of traumatic
seen in the clinic. memories.

Paradoxically, the idea that trauma plays a role in Bibliography

amnesia, while a salient aspect of clinical folklore, is Bootzin, R. R., Kihlstrom, J. F., and Schacter, D. L. (1990). Sleep
not supported by solid empirical evidence. Animal and cognition. Washington, DC: American Psychological Asso-
studies indicate that high levels of emotional arousal ciation.
Eich, E., Macaulay, D., Lowewenstein, R. J., and Dihle, P. H.
stimulate the release of stress hormones that activate (1996). Memory, amnesia, and dissociative identity disorder.
the amygdala, leading to enhanced memory. A review Psychological Science 8, 417422.
of more than sixty longitudinal studies of document- Ellman, S. J., and Antrobus, J. S. (1991). The mind in sleep: Psychology
ed trauma survivors yielded not a single instance of and psychophysiology. New York: Wiley.
amnesia that could not be accounted for by nontrau- Howe, M. L. (2000). The fate of early memories: Developmental science
and the retention of childhood experiences. Washington, DC:
matic infantile and childhood amnesia or organic
American Psychological Association.
factors such as intoxication, anoxia, or head injury. Kapur, N. (1999). Syndromes of retrograde amnesia: A conceptual
Many clinical studies alleging repression of trauma and empirical synthesis. Psychological Bulletin 125, 800825.
also fail to rule out such factors, especially in cases of Kihlstrom, J. F. (1994). One hundred years of hysteria, Dissociation:
infantile and childhood amnesia that occur indepen- Clinical and theoretical perspectives. New York: The Guilford
dently of trauma. Others fail to confirm that the os- Press.
(1996). The trauma-memory argument and recovered
tensibly traumatic event even occurred or misinter-
memory therapy. In K. Pezdek and W. P. Banks, eds.,The re-
pret subjects failure to disclose their trauma as a covered memory/false memory debate. San Diego: Academic Press.
failure to remember it. (1997). Suffering from reminiscences: Exhumed memory,
There are very few corroborated reports of the implicit memory, and the return of the repressed. In M. A.
Conway, ed., Recovered memories and false memories. Oxford:
therapeutic recovery of traumatically lost memories. Oxford University Press.
Most clinicians apparently assume their patients (2001). Dissociative disorders. In P. B. Sutker and H. E.
memories are valid or refrain from challenging them Adams, eds., Comprehensive handbook of psychopathology. New
in order to protect the therapeutic alliance. The York: Plenum.
techniques of memory work used therapeutically to Kihlstrom, J. F., and Barnhardt, T. M. (1993). The self-regulation
of memory: For better and for worse, with and without hypno-
help patients recover traumatic memories also in-
sis, In D. M. Wegner and J. W. Pennebaker, eds., Handbook of
crease the risk of memory distortion and confabula- mental control. Englewood Cliffs, NJ: Prentice-Hall.
tion. There is no evidence that either hypnosis or se- Kihlstrom, J. F., and Schacter, D. L. (2000). Functional amnesia.
dation by barbiturate drugs facilitates the recovery of In F. Boller and J. Grafman, eds., Handbook of neuropsychology.
valid traumatic memories. Whenever recovered Amsterdam: Elsevier.
memories are taken as evidence, whether in the clinic Kopelman, M. D. (1995). The assessment of psychogenic amnesia.
In A. D. Baddeley, B. A. Wilson, and F. N. Watts, eds., Hand-
or the courtroom, it is critical that they be accompa-
book of memory disorders. Chichester, UK: Wiley.
nied by independent, objective, corroborative evi- (1997). Anomalies of autobiographical memory: Retro-
dence. grade amnesia, confabulation, delusional memory, psycho-
genic amnesia, and false memories. In J. D. Read and D. S.
Lindsay, eds., Recollections of trauma: Scientific evidence and clini-
Neuropsychology of Functional Amnesia cal practice. New York: Plenum.
Markowitsch, H. J. (1999). Functional neuroimaging correlates of
Beginning with the discovery of the syphilis spiro- functional amnesia. Memory 7, 561583.
chete, a major goal of psychiatry has been to transfer McNally, R.J. (2002). Remembering Trauma., Cambridge, MA: Har-
syndromes from the functional to the organic catego- vard University Press.
ry as their neural bases are revealed. It seems likely Piper, A. (1993). Truth serum and recovered memories of sex-
that neuroscientific research will pinpoint such causal ual abuse: A review of the evidence. Journal of Psychiatry and
the Law 21, 447471.
relationships for the dissociative amnesias as well. For
Piper, A., Pope, H. G., and Borowiecki, B. S. (2000). Custers last
example, a recent PET study of dissociative fugue stand: Brown, Schefflin, and Whtfields latest attempt to sal-
found no sign of the activity in the right posterior vage dissociative amnesia. Journal of Psychiatry and Law 28
frontal and anterior temporal lobes that usually ac- 149213.
companies recollection of emotionally salient person- Pope, H. G., Oliva, P. S., and Hudson, J. I. (2000). Repressed mem-
al experiences but did find activation in the left fron- ories: B. Scientific status. In D. L. Faigman, D. H. Kaye, M.
J. Saks, and J. Sanders, eds., Modern scientific evidence: The law
tal and temporal regions. Although functional may and science of expert testimony. St. Paul, MN: West Publisher.
eventually prove to be a misnomer, for now the term Reed, G. (1988). The psychology of anomalous experience. Buffalo, NY:
delineates a class of amnesias in which psychological Prometheus.
processes rather than brain insult, injury, or disease Schacter, D. L. (1986). Amnesia and crime: How much do we really
are the immediate causes of the memory failure. know? American Psychologist 41, 286295.
(1987). Implicit memory: History and current status. Jour-
nal of Experimental Psychology: Learning, Memory, and Cognition
AMNESIA, TRANSIENT GLOBAL; HYPNOSIS AND (2001). The seven sins of memory: How the mind forgets and re-
MEMORY members. Boston: Houghton-Mifflin.

Shobe, K. K., and Kihlstrom, J. F. (2002). Interrogative suggestibil- The Lower Boundary for Long-Term Recall
ity and memory work. In M. L. Eisen, J. Quas, and G. S.
Goodman, eds., Memory and suggestibility in the forensic inter-
of Early Experiences
view. Mahwah, NJ: Erlbaum. Maturation of the Central Nervous System
Sullivan, M. D. (1990). Organic or functional? Why psychiatry
needs a philosophy of mind. Psychiatric Annals 20, 271277. Maturation of the human brain begins at concep-
tion, but continues throughout childhood (and be-
John F. Kihlstrom
yond). Although our understanding of the time
Daniel L. Schacter
course of human brain development is not complete,
we do know that many of the brain areas that play a
role in long-term memory are not fully mature during
infancy and early childhood. Thus, although learning
occurs rapidly during this phase of development, the
ability to retain and use information over a lifetime
Do you remember being born? Your first birthday may be precluded by the immaturity of the brain
party? Your first day of school? Despite the signifi- (Campbell and Spear, 1972).
cance of these early experiences, most adults recall lit- Maturation of two areas of the brainin particu-
tle or nothing about them. The absence of autobio- lar, the medial temporal lobe (including the hippo-
graphical memory for events that occurred during campus) and the frontal cortexis thought to play a
infancy and early childhood is commonly referred to particularly important role in the phenomenon of in-
as infantile (or childhood) amnesia. Sigmund Freud fantile amnesia (Bachevalier, 1992). Maturation of
originally identified the phenomenon of infantile am- the hippocampus occurs relatively early in develop-
nesia by asking his patients to describe their earliest ment and may be sufficient to support some of the so-
personal memories in the course of therapy. On the phisticated memory skills exhibited by infants; how-
basis of these patient reports, Freud argued that the ever, maturation of the higher-association areas of
period of infantile amnesia extended into the sixth or the frontal cortex continues well into childhood and
eighth year of life. Freuds most often-cited explana- may be required for the maintenance and retrieval of
tion of infantile amnesia was highly influenced by his memories over the long term (Hayne, Boniface, and
patient population. He believed that memories for Barr, 2000; C. Nelson, 1995).
our infancy and early childhood were stored in pris-
tine condition, but were actively repressed due to The Development of Language
their emotionally and sexually charged content. In When we ask adults to recall their earliest person-
fact, one goal of Freuds psychoanalytic process was al memories, we commonly ask them to provide a ver-
to unlock these hidden memories to allow patients bal report of what they can rememberboth the in-
to come to terms with the traumatic thoughts and ex- structions they are given (tell me about your earliest
periences of their childhood. memory) and their response to those instructions re-
quire sophisticated language skills. Infants and chil-
Subsequent normative studies of adults earliest
dren, on the other hand, typically express their mem-
memories have shown that Freud probably overesti-
ories, by necessity, through nonverbal behaviors.
mated the period of infantile amnesia. There is now
Even once they have acquired conversational lan-
general consensus that adults earliest autobiographi-
guage skills, children still rely primarily on their non-
cal memories are for events that occurred when they
verbal skills to solve tasks that require memory. Fur-
were approximately three to four years of age (Bruce,
thermore, the ability to translate early, preverbal
Dolan, and Phillips-Grant, 2000; Dudycha and Dudy-
experiences into language is extremely limited, if not
cha, 1941; Mullen, 1994; Sheingold and Tenney,
impossible (Simcock and Hayne, 2002). Although an
1982; Waldfogel, 1948) or even slightly younger
early preverbal memory may be reflected in some as-
(MacDonald, Uesiliana, and Hayne, 2000; Usher and
pect of an adults behavior (Newcombe, Drummey,
Neisser, 1993). Furthermore, normative studies of
Fox, Lie, and Ottinger-Alberts, 2000), he or she will
adults earliest memories have failed to provide any
be unable to provide a verbal report of the original
empirical evidence in support of Freuds repression
experience. In this way, language development is an-
model (Pillemer and White, 1989).
other rate-limiting step in the offset of infantile am-
Thus, the fundamental question remains: Why is nesia.
it that we have little or no recollection of events that
occurred during our infancy and early childhood? Al-
though repression does not provide an adequate ex-
Beyond the Basic Ingredients: The
planation for the phenomenon, empirical studies Emergence of Autobiographical Memory
point to a number of other factors that might account The basic ingredients for long-term verbal mem-
for infantile amnesia (Howe and Courage, 1993). ory are in place by the end of the second year of life.

For example, children as young as two and a half can ory for a particular experience (e.g., Remember
provide a verbal report of an event that occurred when we went to the zoo last year?). Furthermore, an
eighteen months earlier (Fivush and Hammond, adults account of an event may augment the childs
1990). Despite this, most adults can recall nothing memory for the same experience, yielding a richer,
about events that occurred prior to their third or more integrated (and thus more memorable) repre-
fourth birthday, and even those memories are few sentation. The frequency and content of these conver-
and far between and are significantly less vivid or de- sations ultimately shape the number and clarity of our
tailed than events that occurred later in childhood. earliest recollections (MacDonald, Uesiliana, and
How can we account for this aspect of infantile amne- Hayne, 2000; Mullen, 1994).
sia? In conclusion, no single explanation of infantile
A Framework for Organization amnesia can account for all of the available data. In-
stead, it is likely that brain maturation and language
Many lines of research have shown that memory
acquisition define the lower limit for our earliest rec-
is enhanced when the to-be-remembered information
ollections, but the number and quality of memories
can be organized according to some cognitive frame-
that actually survive will be determined by the way in
work, or schema. Whether the material to be remem-
which those memories have been structured and or-
bered consists of stories, baseball statistics, or lists of
ganized. Conversations about the past during early
words, it is learned and retrieved most effectively by
childhood are a driving force in this process.
people who can systematically organize the material
by relating it to an existing framework of knowledge.
Without this kind of schematic organization, recall is See also: CHILDREN, DEVELOPMENT OF MEMORY IN;
only partial and fragmentary, if it occurs at all. CODING PROCESSES: ORGANIZATION OF
This same principle also applies to the recall of MEMORIES; GUIDE TO THE ANATOMY OF THE
autobiographical information. When we think or talk BRAIN; NATURAL SETTINGS, MEMORY IN; PROSE
about our past experiences, for example, we typically RETENTION
recall that information in chunks that reflect mile-
stones in our lives: graduation from high school, at-
tending college, getting married, the birth of our chil- Bibliography
dren and grandchildren, and so on. Our schema for Bachevalier, J. (1992). Cortical versus limbic immaturity: Relation-
ship to infantile amnesia. In M. R. Gunnar and C. A. Nelson,
these stages in our lives helps us to retrieve individual eds., Developmental behavioral neuroscience, pp. 129153. Hills-
memories that occurred during each stage. Further- dale, NJ: Erlbaum.
more, thinking or talking about experiences from a Bruce, D., Dolan, A., and Phillips-Grant, K. (2000). On the transi-
particular stage often reminds us of other events that tion from childhood amnesia to the recall of personal memo-
occurred during the same stage even if those events ries. Psychological Science 11, 360364.
Campbell, B. A., and Spear, N. E. (1972). Ontogeny of memory.
were greatly separated in time. Psychological Review 79, 215236.
Children, on the other hand, do not organize Dudycha, G. J., and Dudycha, M. M. (1941). Childhood memories:
their memories on the basis of milestones that reflect A review of the literature. Psychological Review 38, 668682.
Fivush, R., and Hammond, N. R. (1990). Autobiographical memo-
important stages of development or pivotal life ry across the preschool years: Towards reconceptualizing
events. Because these early experiences are not linked childhood amnesia. In R. Fivush and J. A. Hudson, eds.,
to the same schemata that adults use when they at- Knowing and remembering in young children, pp. 223248. New
tempt to retrieve them, individual memories become York: Cambridge University Press.
difficult, if not impossible to recall (Pillemer and Fivush, R., and Reese, E. (1992). The social construction of autobi-
ographical memory. In M. A. Conway, D. C. Rubin, H. Spinn-
White, 1989). ler, and W. A. Wagenaar, eds., Theoretical perspectives on autobi-
ographical memory, pp. 115132. Dordrecht: Kluwer Academic
Parent-Child Conversations about the Past Publishers.
Childrens emerging ability to organize and re- Freud, S. (1960). Three essays on the theory of sexuality: 2. Infan-
count their life history is shaped through conversa- tile sexuality. In The complete psychological writings of Sigmund
Freud, vol. 7, pp. 173206. London: Hogarth. The original
tions about the past that occur in the context of their
definition of infantile amnesia. First published in 1905.
family (Fivush and Reese, 1992; Hudson, 1990; K. Hayne, H., Boniface, J., and Barr, R. (2000). The development of
Nelson, 1993). In the course of reminiscing about declarative memory in human infants: Age-related changes in
mutually experienced events with parents or other deferred imitation. Behavioral Neuroscience 114, 7783.
significant social partners, children acquire the sche- Howe, M., and Courage, M. (1993). On resolving the enigma of in-
fantile amnesia. Psychological Bulletin 113, 305326.
ma that are necessary to catalog their autobiographi-
Hudson, J. A. (1990). The emergence of autobiographical memory
cal memories. Additionally, conversations about the in mother-child conversation. In R. Fivush and J. A. Hudson,
past allow children the opportunity to practice using eds., Knowing and remembering in young children, pp. 166196.
another persons language to retrieve their own mem- New York: Cambridge University Press.

MacDonald, S., Uesiliana, K., and Hayne, H. (2000). Cross-cultural learning word pairs), and recognition memory. His
and gender differences in childhood amnesia. Memory 8, 365 impairment is global in nature: information received
Mullen, M. K. (1994). Earliest recollections of childhood: A demo-
from all sensory modalities is affected, and both ver-
graphic analysis. Cognition 52, 5579. bal and nonverbal (e.g., spatial) memory are im-
Nelson, C. (1995). The ontogeny of human memory: A cognitive paired. H.M. also evidences deficient recall of remote
neuroscience perspective. Developmental Psychology 31, 723 memories antedating the surgery (retrograde amne-
738. sia). For instance, H.M. could not recall the death of
Nelson, K. (1993). The psychological and social origins of autobio-
graphical memory. Psychological Science 4, 714.
a favorite uncle who had died three years prior to the
Newcombe, N. S., Drummey, A. B., Fox, N. A., Lie, E., and Ot- surgery. Moreover, nearly all of his personal memo-
tinger-Alberts, W. (2000). Remembering early childhood: ries are from the age of sixteen or earlier (Sagar,
How much, how, and why (or why not). Current Directions in Cohen, Corkin, and Growdon, 1985).
Psychological Science 9, 5558.
Pillemer, D. B., and White, S. H. (1989). Childhood events recalled Despite the severity of his amnesia, H.M. exhibits
by children and adults. In H. W. Reese, ed., Advances in child normal attention, language, and general intellectual
development and behavior, Vol. 21, pp. 297340. Orlando, FL:
abilities. His performance on tasks of immediate, or
Academic Press.
Sheingold, K., and Tenney, Y. J. (1982). Memory for a salient working, memory is relatively preserved. That is, he
childhood event. In U. Neisser, ed., Memory observed, pp. 201 can temporarily store and maintain information over
212. San Francisco: Freeman. a brief interval, such as that required when rehearsing
Simcock, G., and Hayne, H. (2002). Breaking the barrier? Children a telephone number. But if he is distracted or pre-
fail to translate their preverbal memories into language. Psy-
vented from continually rehearsing the material, the
chological Science 13, 225231.
Usher, J. A., and Neisser, U. (1993). Childhood amnesia and the information is forgotten. H.M. exhibits preservation
beginnings of memory for four early life events. Journal of Ex- of some long-term memory functions, such as skill
perimental Psychology: General 122, 155165. and habit learning, where learning is expressed as en-
Waldfogel, S. (1948). The frequency and affective character of hanced task performance. The detailed examination
childhood memories. Psychological Monographs 62, 139.
of H.M.s amnesia has served as a milestone in the
Ulric Neisser quest to elucidate the functional and neuroanatomi-
Revised by Harlene Hayne cal basis of memory and has spawned decades of re-
search into the memory processes that are impaired
and preserved in amnesia.

Etiologies of Amnesia
Organic amnesia is a neurological disorder character-
ized by a dense impairment of memory in the context Although H.M. still serves as a benchmark for
of normal intelligence and other preserved mental characterizing global amnesia, it has also become
abilities. Investigations of patients with this disorder clear that the disorder is composed of a number of
have enhanced the understanding of the psychologi- different patterns of memory loss that may be linked
cal processes involved in learning and remembering, to distinct etiologies and associated patterns of brain
as well as the brain organization of human memory. damage. Organic amnesia has been associated with a
wide range of medical conditions including vascular
Much of the current interest in memory and brain accidents (i.e., strokes), ischemia (e.g., loss of blood
function finds its origin in the study of patient H.M., flow to the brain), anoxia (i.e., loss of oxygen to the
a man who, in 1953, underwent surgery for treatment brain), viral infections, and Wernicke-Korsakoff syn-
of refractory seizures (Scoville and Milner, 1957). The drome. As with H.M., patients with these medical con-
surgery involved bilateral resection of a large portion ditions typically show preserved attention, working
of the medial temporal region, which includes the memory, and general intellectual abilities.
amygdala, hippocampus, and hippocampal gyrus. Al-
though the surgery was successful in substantially re- Not all forms of amnesia are permanent. For ex-
ducing H.M.s seizures, the procedure produced a ample, head injury can cause transient and selective
pervasive impairment of memory (Milner, Corkin, memory impairment. Anterograde amnesia following
and Teuber, 1968). Since the time of his surgery at head trauma can last minutes, days, or even weeks.
the age of twenty-seven, H.M. has been unable to con- Depending on whether the trauma is mild or severe,
sciously learn and remember new information (an- patients may completely regain their learning ability
terograde amnesia). For example, thirty minutes or may suffer long-lasting and sometimes permanent
after eating lunch, H.M. cannot recall what he ate for impairment. Retrograde amnesia may also occur, and
lunch, nor can he recall if in fact he ate lunch at all. the temporal extent of retrograde amnesia is often
He exhibits severe impairment on laboratory tests of correlated with the severity of anterograde amnesia.
word and picture recall, cued-word learning (e.g., Quite severe memory problems also occur after elec-

troconvulsive therapy (ECT), a procedure sometimes disorder resulting from the convergent effects of
prescribed for treatment of severe depression. How- chronic alcohol abuse and malnutrition. Studies of
ever, ECT-induced amnesia is usually transient, and postmortem brain tissue by Victor, Adams, and Col-
extensive recovery of memory occurs with time. lins (1989) have revealed bilateral damage involving
the dorsomedial nucleus of the thalamus and a sub-
thalamic nucleus called the mamillary bodies. Similar
The Anatomy of Memory pathology can also occur in nonamnesic alcoholics,
Analysis of the locus of brain lesions in patients but what distinguishes patients with Wernicke-
such as H.M. has underscored the importance of the Korsakoff syndrome is that they also show neuronal
medial temporal region, particularly the hippocam- loss in anterior thalamic nuclei (Harding, Halliday,
pus and adjacent regions, in new learning. Further in- Caine, and Kril, 2000). Cortical atrophy (i.e., brain-
sight into the prominent role of the hippocampus in cell loss) and cerebellar damage are also often ob-
memory comes from the study of patient R.B. by Zola- served. On neuropsychological tests Wernicke-
Morgan, Squire, and Amaral (1986). R.B. became am- Korsakoff patients evidence both anterograde amne-
nesic in 1978 when he suffered an ischemic episode sia and retrograde amnesia. In addition, these pa-
during open-heart surgery. Neuropsychological as- tients may exhibit attention and problem-solving
sessment of R.B. revealed moderate level antero- impairments, as well as impaired insight. These addi-
grade amnesia alongside mild retrograde amnesia. In tional difficulties may occur as a result of cortical atro-
1983, R.B. suffered a fatal cardiac arrest and his brain phy, especially in prefrontal areas.
was brought to autopsy. Examination of his brain re-
vealed that R.B. had sustained discrete bilateral brain
damage restricted to a portion of the hippocampus Declarative Memory in Amnesia
called the CA1 subfield. Several other cases of amne- Globally amnesic patients are severely impaired
sia resulting from ischemia have come to autopsy at consciously and intentionally remembering infor-
since then (Rempel-Clower, Zola, Squire, and Ama- mation. This deficit encompasses the acquisition,
ral, 1996). These cases have revealed that more ex- long-term retention, and retrieval of both personally
tensive damage to the hippocampal formation pro- experienced events (i.e., episodic memory) and im-
duces more severe anterograde amnesia, as well as personal information (i.e., semantic memory). Collec-
extensive retrograde amnesia for memories predat- tively, these forms of memory are referred to as declar-
ing the brain injury up to fifteen years or more. In ative memory.
many patients, lesions extend beyond the hippocam- Episodic memory enables individuals to remem-
pal formation to include adjacent regions (i.e., en- ber experiences from their personal past (e.g., re-
torhinal and perirhinal). This typically leads to dense membering what they ate for breakfast this morning).
anterograde and retrograde amnesia. Importantly, Episodic memories contain a multitude of sources of
comparisons between patients with restricted versus information including perceptual, conceptual, and
extensive medial temporal lesions suggest that the emotional components. Episodic memories are not
hippocampal formation and adjacent regions make stored in isolation, but are placed within a context of
qualitatively different contributions to memory. It has personally relevant information. A pervasive deficit in
been suggested that one neural circuit centered in the episodic memory is dramatically exemplified by glob-
hippocampus supports recollection (i.e., the con- ally amnesic patients. In the clinic, episodic memory
scious or intentional retrieval of past experiences) is assessed by tests of recall and recognition. Because
whereas another neural circuit centered in adjacent these tasks require intentional retrieval of recent ex-
perirhinal regions mediates overall feelings of famil- periences, they are referred to as tests of explicit memo-
iarity (i.e., the subjective sense that something was en- ry. Although both recall and recognition require in-
countered previously) (Aggleton and Brown, 1999; tentional retrieval, their processing demands are not
Brown and Aggleton, 2001). identical. Recognition memory is based on two dis-
Damage to another region of the brain, the dien- tinct memory processes: recollection, an effort-
cephalic midline, can also produce organic amnesia. dependent process by which information is deliber-
This brain region includes various midline thalamic ately brought to mind (e.g., remembering what book
nuclei (nuclei are groups of neurons), as well as sub- you read last night); and familiarity, a facilitation of
thalamic nuclei. These neurons serve as relay groups, or fluency in stimulus processing that results from
sending and receiving projections to numerous parts prior experience with that stimulus (e.g., the sense of
of the brain, including the medial temporal region. familiarity when seeing a person on the bus, even
The best-studied cases of amnesia resulting from though you may not remember when you last saw him
damage to diencephalic midline structures are pa- or her). By contrast, recall depends largely on recol-
tients with Wernicke-Korsakoff syndrome, an amnesic lection. Most globally amnesic patients evidence defi-

cits both on recall and on recognition tests, but their amnesic individuals. Strikingly, however, H.M. had
recall performance is typically worse than their recog- no awareness of having practiced the task.
nition performance (Giovanello and Verfaellie,
2001). This finding suggests that recollection may be Another form of nondeclarative memory is repe-
more severely disrupted in amnesia than is familiarity tition priming, the facilitation in performance in-
(Yonelinas et al., 1998). duced by previous exposure to stimuli. A typical prim-
ing experiment is composed of a study phase
Whereas episodic memory is concerned with per- followed by a test phase. During the study phase,
sonally experienced events, semantic memory refers participants are exposed to a list of words or pictures.
to the acquisition and long-term retention of generic For example, participants might see a list containing
factual information. Semantic knowledge encompass- the word apricot. During the subsequent test phase,
es a wide range of information, including facts about participants are asked to perform a seemingly unre-
the world (e.g., the knowledge that Rome is the capi- lated task. For example, they may be asked to identify
tal of Italy), the meanings of words and concepts (e.g., briefly flashed words or to generate as many words as
an understanding of the concept Website), and the possible when cued with the semantic category
names attached to objects and people (e.g., the fruit. Priming is measured as the change in accura-
knowledge that William Shakespeare wrote Hamlet). cy or the bias in task performance induced by recent
Studies of H.M. (Gabrieli, Cohen, and Corkin, 1988) exposure to task stimuli (e.g., enhanced accuracy at
and other amnesic patients (Verfaellie, Reiss, and identifying the word apricot or enhanced likelihood of
Roth, 1995) suggest that semantic learning is im- generating the word apricot), compared to a test con-
paired in amnesia. For example, H.M. has been un- dition in which apricot did not appear on the prior
able to acquire any new vocabulary words that entered study list.
the language since his surgery. Other amnesic pa-
tients, however, appear able to acquire some new Priming in amnesia has been assessed using tasks
semantic information, and occasionally semantic that require analysis of the perceptual attributes of a
learning is quite good despite a patients severe im- stimulus (perceptual priming), such as identification
pairment in episodic memory (Verfaellie, 2000). For of briefly presented words or pictures and speeded
example, Jon, a young patient who became amnesic reading of words. Priming has also been assessed
following an anoxic episode at birth, has been able to using tasks that require analysis of the meaning of a
acquire a considerable amount of semantic knowl- stimulus (conceptual priming), such as category
edge, as demonstrated by his ability to attend main- exemplar generation and production of word asso-
stream schools, despite his pronounced amnesia for ciates. Amnesic patients show intact priming on both
day-to-day events, such as remembering conversa- perceptual and conceptual priming tasks, suggesting
tions or television programs (Baddeley, Vargha- that these forms of unaware memory do not depend
Khadem, and Mishkin, 2001). The degree to which on the medial temporal regions implicated in amne-
semantic learning is still possible in amnesic patients sia. The neural basis of conceptual priming is not well
likely depends on the extent of lesion in the temporal understood, but findings of impaired visual perceptu-
lobe. al priming in patients with occipital lesions suggest
that perceptual priming is mediated by posterior cor-
tical areas that process modality-specific information
Nondeclarative Memory in Amnesia (Verfaellie and Keane, 1997).
One of the most striking findings about amnesia
is that, despite their severe impairment in conscious,
intentional retrieval of information (declarative Remote Memory in Amnesia
memory), amnesic patients can acquire several forms Globally amnesic patients evidence deficits in
of memory normally. These forms of memory are col- memory for events and for facts predating the onset
lectively referred to as nondeclarative memory. One of their amnesia, although the extent of retrograde
form of nondeclarative memory is procedural learn- memory loss varies from patient to patient. Retro-
ing, the ability to acquire new perceptual and motor grade amnesia tends to follow Ribots law, which
skills on the basis of repeated practice (e.g., learning states that memory for the recent past is more severe-
to ride a bicycle). Studies of H.M. were among the ly affected than memory for the remote past. This
first to establish the preservation of procedural learn- pattern of retrograde memory loss provides clues
ing in patients with global amnesia. For instance, dur- about the mechanisms and time course involved in
ing one task, H.M. had to learn how to keep a metal the storage and retrieval of new memories. Although
stylus in contact with a revolving disc. He learned to the medial temporal region plays a critical role in the
adapt his motor movement to the movement of the permanent laying down of information, these brain
disc and gradually got better at the task, just like non- areas are not the ultimate repositories for new memo-

ries. Storage of new memories requires interaction Brown, M. W., and Aggleton, J. P. (2001). Recognition memory:
between medial temporal and neocortical (outer What are the roles of the perirhinal cortex and hippocampus?
Nature Review Neuroscience 2, 5161.
brain) regions. The hippocampal formation receives Gabrieli, J. D. E., Cohen, N. J., and Corkin, S. (1988). The im-
information from a number of neocortical sites and paired learning of semantic knowledge following bilateral
binds together the spatial, perceptual, conceptual, medial temporal-lobe resection. Brain and Cognition 7, 525
and emotional components of an experience. Subse- 539.
quent attempts to retrieve this experience cause the Giovanello, K. S., and Verfaellie, M. (2001). The relationship be-
tween recall and recognition in amnesia: Effects of matching
hippocampus to reactivate the neocortical sites that recognition between patients with amnesia and controls.
contain the information. Reactivation of the neocorti- Neuropsychology 15, 444451.
cal sites leads to a strengthening of the connections Harding, A., Halliday, G., Caine, D., and Kril, J. (2000). Degenera-
between these sites and, over time, allows memories tion of anterior thalamic nuclei differentiates alcoholics with
to be retrieved without assistance from the hippocam- amnesia. Brain 123 (1), 141154.
Milner, B., Corkin, S., and Teuber, H.-L. (1968). Further analysis
pal formation (Alvarez and Squire, 1994). Conse- of the hippocampal amnesia syndrome: Fourteen-year follow-
quently, in patients with damage to medial temporal up study of H.M. Neuropsychologia 6, 215234.
or connected diencephalic structures, information Rempel-Clower, N. L., Zola, S. M., Squire, L. R., and Amaral, D.
that has not been fully consolidated is vulnerable, G. (1996). Three cases of enduring memory impairment after
whereas fully consolidated older memories can still be bilateral damage limited to the hippocampal formation. Jour-
nal of Neuroscience 16, 5,2335,255.
retrieved. Sagar, H. J., Cohen, N. J., Corkin, S., and Growdon, J. H. (1985).
Dissociations among processes in remote memory. In D. S.
Olton, E. Gamzu, and S. Corkin, eds., Memory dysfunctions: An
Conclusion integration of animal and human research from preclinical and clini-
Neuropsychological investigations of patients cal perspectives, Vol. 4, pp. 533535. New York: New York
Academy of Sciences.
with organic amnesia have contributed greatly to the
Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
understanding of memory processes and the brain re- bilateral hippocampal lesions. Journal of Neurology, Neurosur-
gions that mediate them. Damage to medial temporal gery, and Psychiatry 20, 1121.
regions or diencephalic midline structures produces Verfaellie, M. (2000). Semantic learning in amnesia. In Handbook
a pervasive amnesic disorder in which conscious or of neuropsychology, 2nd edition, Vol. 2: Memory and its disorders,
ed. L. S. Cermak, pp. 335354. Amsterdam: Elsevier Science.
declarative memory is severely impaired. In contrast,
Verfaellie, M., and Keane, M. M. (1997). The neural basis of
various forms of nondeclarative memory such as pro- aware and unaware forms of memory. Seminars in Neurology
cedural memory and repetition priming are pre- 17, 7583.
served, suggesting that nondeclarative memory is not Verfaellie, M., Reiss, L., and Roth, H. (1995). Knowledge of new
mediated by the medial temporal region implicated English vocabulary in amnesia: An examination of premor-
bidly acquired semantic memory. Journal of the International
in amnesia. These theoretical advances provide the
Neuropsychological Society 1, 443453.
opportunity for more sophisticated assessment of pa- Victor, M., Adams, R., and Collins, G. (1989). The Wernicke-
tients with memory impairments and may lead to the Korsakoff syndrome and related neurologic disorders due to alcohol-
development of new rehabilitation techniques that ism and malnutrition. Philadelphia: F. A. Davis.
capitalize on preserved forms of memory. Yonelinas, A. P., Kroll, N. E., Dobbins, I., Lazzara, M., and Knight,
R. T. (1998). Recollection and familiarity deficits in amnesia:
Convergence of remember-know, process dissociation, and
receiver operating characteristic data. Neuropsychology 12,
LOSS; EPISODIC MEMORY; IMPLICIT MEMORY; Zola-Morgan, S., Squire, L. R., and Amaral, D. G. (1986). Human
MEMORY SPAN; PROCEDURAL LEARNING: amnesia and the medial temporal region: Enduring memory
HUMANS; RIBOT, THODULE; SEMANTIC impairment following bilateral lesion limited to the field CA1
MEMORY: COGNITIVE ASPECTS; SEMANTIC of the hippocampus. Journal of Neuroscience 6, 2,9502,967.
Arthur P. Shimamura
Revised by Kelly Sullivan Giovanello and Mieke Verfaellie
Aggleton, J. P., and Brown, M. W. (1999). Episodic memory, amne-
sia, and the hippocampal-anterior thalamic axis. Behavioral
and Brain Sciences 22, 425489. AMNESIA, TRANSIENT GLOBAL
Alvarez, P., and Squire, L. R. (1994). Memory consolidation and
the medial temporal lobe: A simple network model. Proceed- Transient global amnesia (TGA) is a benign neuro-
ings of the National Academy of Sciences of the United States of Amer- logical condition in which the prominent deficit is a
ica 91, 7,0417,045. temporary organic amnesic syndrome. The episode
Baddeley, A., Vargha-Khadem, F., and Mishkin, M. (2001). Pre-
served recognition in a case of developmental amnesia: Impli- of TGA is stereotyped. It usually begins suddenly,
cations for the acquisition of semantic memory? Journal of Cog- lasts for at least several hours, and resolves gradually
nitive Neuroscience 13, 357369. over several hours to a day. Careful examination dur-

ing an episode of TGA shows that the patient has a tained by well-studied patients with chronic organic
relatively isolated amnesic syndrome. Vision, hear- amnesia. The severity of the anterograde amnesia ap-
ing, sensation, strength, and coordination are nor- peared to correlate with the time since onset of TGA.
mal. Language, spatial abilities, and general intellec- There was no evidence for material-specific, or par-
tual function also are normal, and the TGA patient tial, amnesiano TGA patient had a significant dis-
can repeat a list of numbers or words. In contrast, the parity between the degree of anterograde amnesia for
patient can recall little of any verbal or nonverbal ma- verbal and nonverbal material. The patients also had
terial presented minutes before. The TGA patient a temporally graded retrograde amnesia (i.e., inabili-
often repeats the same question many times because ty to recall events that occurred before the onset of
of an inability to remember the answer that was just amnesia) covering at least twenty years prior to TGA
given. Frequently repeated questions include, Is onset. Again, their test scores were similar to the
there something the matter with me? and Whats scores of many well-studied patients with chronic or-
wrong, have I had a stroke? During TGA the patient ganic amnesia. The retrograde amnesia was patchy
has a patchy loss of recall for events dating from sev- all patients were able to recall some events that oc-
eral hours to many years before the attack. Older curred within the time interval affected by retrograde
memories are spared, and the patient does not lose amnesia. Indeed, during TGA some memories were
personal identity. TGA patients examined carefully recalled, albeit incompletely, that were less than one
after the episode are normal except for an inability to to two months old. The temporally graded retrograde
recall the episode. amnesia was similar for both public and personal
events. During the episode, TGA patients performed
TGA generally occurs in persons over the age
normally on almost all other formal neuropsychologi-
fifty, and 75 percent of patients are fifty to sixty-nine
cal tests.
years old. Men and women are affected with nearly
equal frequency. The estimated incidence is 5.2 per TGA is likely caused by temporary dysfunction of
100,000 per year for persons of all ages and 23.5 per either bilateral medial temporal lobe structures, in-
100,000 per year for persons older than fifty. A third cluding field CA1 of the hippocampus and adjacent,
to a half of TGA attacks are precipitated by physical anatomically related, structures; or bilateral medial
or psychological stress, including strenuous exertion, diencephalic structures, including the dorsomedial
sexual intercourse, intense emotion, pain, exposure nucleus of the thalamus, the mammillothalamic tract,
to intense heat or cold, and minor or major medical and the mammillary bodies. The exact cause of TGA
procedures. TGA has a recurrence rate of 3 to 5 per- is not known, although it appears to be a benign con-
cent per year for at least five years after the initial epi- dition that requires no medical treatment.
sode. However, the patient with TGA does not appear
to have an increased risk of developing significant
permanent memory deficit or other cognitive dys- Bibliography
function, and TGA patients have an incidence of sub- Hodges, J. R. (1998). Unraveling the enigma of transient global
sequent stroke equal to the incidence of a comparable amnesia. Annals of Neurology 43, 151153.
population. Kritchevsky, M. (1987). Transient global amnesia: When memory
temporarily disappears. Postgraduate Medicine 82, 95100.
Few laboratory abnormalities are associated with (1989). Transient global amnesia. In F. Boller and J. Graff-
TGA. Blood, urine, and spinal fluid examination, man, eds., Handbook of neuropsychology, Vol. 3, pp. 167182.
electrocardiogram, electroencephalogram, and brain Amsterdam: Elsevier.
computerized tomography are normal. Some investi-
Mark Kritchevsky
gators have noted abnormalities, particularly in one
or both medial temporal lobes, on single-photon
emission computerized tomography (SPECT) brain
scans and on diffusion-weighted magnetic resonance
imaging (MRI) brain scans performed during the epi-
sode of TGA, which resolve following the episode. AMYGDALA
Only a small number of patients have been exam- See: GENETIC SUBSTRATES OF MEMORY:
ined with formal neuropsychological tests during AMYGDALA; GUIDE TO THE ANATOMY OF THE
TGA. Of these, eleven were studied with a single bat-
tery of memory tests that also had been given to pa-
tients with chronic organic amnesia. All of the eleven
patients had severe anterograde amnesia (i.e., inabili-
ty to learn new material) for verbal and nonverbal ma-
terial, and their test scores were similar to scores ob- See: COMPARATIVE COGNITION
APLYSIA: Classical Conditioning and Operant Conditioning 33

APLYSIA the salient event with a fixed latency, then the animal
learns that the stimulus can serve as a predictor.
[Since the mid-1960s, the marine mollusk Aplysia has
proved to be an extremely useful model system for gaining Behavioral Studies
insights into the neural and molecular mechanisms of simple A tactile or electrical stimulus delivered to the si-
forms of memory. Indeed, the pioneering discoveries of Eric phon results in a reflex withdrawal of the gill and si-
Kandel using this animal were recognized by his receipt of phon, which presumably serves the defensive role of
the Nobel Prize in physiology or medicine in 2000. A num- protecting these sensitive structures from potentially
ber of characteristics make Aplysia well suited to the exami- harmful stimuli. The reflex exhibits several simple
nation of the molecular, cellular, morphological, and net- forms of learning, including habituation, sensitiza-
work mechanisms underlying neuronal modifications tion, and classical conditioning. In studies of aversive
(plasticity) and learning and memory. The animal has a rel- classical conditioning, the conditioned stimulus (CS)
atively simple nervous system with large, individually identi- is a brief, weak tactile stimulus to the siphon, which
fiable neurons that are accessible for detailed anatomical, by itself produces a small siphon withdrawal. The un-
biophysical, biochemical, and molecular studies. Neurons conditioned stimulus (US) is a short-duration, strong
and neural circuits that mediate many behaviors in Aplysia (noxious) electric shock to the tail, which by itself pro-
have been identified. In several cases, these behaviors have duces a large withdrawal of the siphon (the uncondi-
been shown to be modified by learning. Moreover, specific tioned response, UR). After repeated pairings, the
loci within neural circuits where modifications occur during ability of the CS to produce siphon withdrawal (the
learning have been identified, and aspects of the cellular conditioned response, CR) is enhanced beyond that
mechanisms underlying these modifications have been ana- produced by presentations of the US alone (sensitiza-
lyzed. The three entries that follow review several aspects of tion control) or of explicitly unpaired or random pre-
research on Aplysia. CLASSICAL CONDITIONING AND OP- sentations of the CS and the US (Carew, Walters, and
ERANT CONDITIONING describes several of the behaviors of Kandel, 1981). The conditioning persists for as long
Aplysia that exhibit these associative forms of learning and as four days. Thomas J. Carew and colleagues (Carew,
the progress that has been made in examining the underlying Hawkins, and Kandel, 1983) also found that this re-
mechanisms. MOLECULAR BASIS OF LONG-TERM flex exhibits differential classical conditioning. Dif-
SENSITIZATION describes the elucidation of second- ferential classical conditioning can be produced by
messenger cascades and the roles of specific genes and pro- delivering one CS to the siphon and another to the
teins in the induction and maintenance of this example of mantle region. One CS is paired with the US (the
nonassociative learning. DEVELOPMENT OF PROCESSES CS+) and the other is explicitly unpaired (the CS-).
UNDERLYING LEARNING describes the ways in which it has As in the previous studies, the US is an electric shock
been possible to identify and dissociate multiple components delivered to the tail. After conditioning, the CS+ will
of nonassociative learning on both behavioral and cellular produce more withdrawal than the CS-.
levels. For entries on other invertebrates that have proved
Other behaviors of Aplysia can also be classically
useful for examining mechanisms of learning and memory,
conditioned. For example, feeding behavior can be
classically conditioned with an appetitive protocol
(Lechner, Baxter, and Byrne, 2000).
Neural Mechanisms of Aversive Classical
Conditioning in Aplysia
CLASSICAL CONDITIONING AND A cellular mechanism called activity-dependent
OPERANT CONDITIONING neuromodulation contributes to associative learning
in Aplysia (Hawkins, Abrams, Carew, and Kandel,
The simple nervous system and the relatively large 1983; Walters and Byrne, 1983; Antonov, Antonova,
identifiable neurons of the marine mollusk Aplysia Kandel, and Hawkins, 2001). A general cellular
provide a useful model system in which to examine scheme of activity-dependent neuromodulation is il-
the cellular mechanisms of two forms of associative lustrated in Figure 1. Two sensory neurons (SN1 and
learning: classical conditioning and operant (instru- SN2), which constitute the pathways for the condi-
mental) conditioning. tioned stimuli (CS+ and CS-), make weak subthresh-
old connections to a motor neuron. Delivering a rein-
forcing or unconditioned stimulus (US) alone has two
Classical Conditioning effects. First, the US activates the motor neuron and
Classical conditioning occurs when an animal produces the unconditioned response (UR). Second,
learns to associate a typically neutral stimulus with a the US activates a diffuse modulatory system that
later salient event. If the neutral stimulus precedes nonspecifically enhances transmitter release from all
34 APLYSIA: Classical Conditioning and Operant Conditioning

Figure 1 dependent protein kinase; the subsequent protein

phosphorylation leads to a modulation in several
properties of the sensory neurons. These changes in
the sensory neuron include modulation of membrane
conductances and other processes, which facilitate
synaptic transmission. This facilitation results in the
increased activation of the motor neuron and, thus,
sensitization of the reflex. The pairing specificity of
the associative conditioning is due, at least in part, to
an increase in the level of cAMP beyond that pro-
duced by serotonin alone (Abrams and Kandel, 1988;
Ocorr, Walters, and Byrne, 1985). The influx of Ca+2
associated with the CS (spike activity) amplifies the
US-mediated modulatory effect by interacting with a
Ca+2-sensitive component of the adenylyl cyclase
(Abrams and Kandel, 1988; Schwartz et al., 1983). A
critical role for Ca+2-stimulated cyclase is also sug-
gested by studies of Drosophila showing that the ad-
General model of activity-dependent neuromodulation. A. enylyl cyclase of a mutant deficient in associative
Learning: Shading indicates paired activity. A motivationally learning exhibits a loss of Ca+2/calmodulin sensitivity.
potent reinforcing stimulus (US) activates a motor neuron to
There are contributions to the plasticity of the
produce the unconditioned response (UR) and a facilitatory
synapse occurring in the motor neuron, as well (Mur-
neuron (FN) or modulatory system that regulates the strength
of the connection between sensory neurons (SN1 and SN2) and
phy and Glanzman, 1997; Bao, Kandel, and Hawkins,
the motor neuron. Increased spike activity in one sensory 1998). The postsynaptic membrane of the motor neu-
neuron (SN1) immediately before the modulatory signal ron contains NMDA receptors. If these receptors are
amplifies the degree and duration of the modulatory effects, blocked, then the associative modification of the syn-
perhaps through the Ca+2 sensitivity of the modulatory evoked apse is disrupted. NMDA receptors require concur-
second messenger, with possible contributions from the rent delivery of glutamate and depolarization in
postsynaptic neuron. The unpaired sensory neuron (SN2) does order to allow the entry of calcium. Activity in the sen-
not show an amplification of the modulatory effects. B. sory neuron (CS) provides the glutamate and the US
Memory: The amplified modulatory effects cause long-term depolarizes the cell. The subsequent increase in intra-
increases in transmitter release and/or excitability of the paired
cellular Ca+2 putatively releases a retrograde signal
neuron, which in turn strengthens the functional connection
from the postsynaptic cell to the presynaptic termi-
between the paired sensory neuron (SN1) and the motor
neuron. The associative enhancement of synaptic strength
nal. This retrograde signal would then act to further
represents the conditioned response (CR). enhance the cAMP cascade in the sensory neuron.
An important conclusion is that this mechanism
for associative learning is an elaboration of a process
the sensory neurons. This nonspecific enhancement already in place that mediates sensitization, a simpler
contributes to sensitization. Temporal specificity, form of learning. This finding raises the interesting
characteristic of associative learning, occurs when possibility that even more complex forms of learning
there is pairing of the CS (spike activity in SN1) with may use simpler forms as building blocks, an idea that
the US, causing a selective amplification of the modu- has been suggested by some psychologists for many
latory effects in SN1. Unpaired activity does not am- years but one that until the early years of the twenty-
plify the effects of the US in SN2. The amplification first century has not been testable at the cellular level.
of the modulatory effects in SN1 leads to an enhance-
ment of the ability of SN1 to activate the motor neu- Operant Conditioning
ron and produce the CR.
Operant conditioning is a process by which an an-
As discussed in the entry Molecular Basis of imal learns the consequences of its own behavior. In
Long-Term Sensitization, experimental analyses of an operant-conditioning paradigm, the delivery of a
sensitization of defensive reflexes in Aplysia have reinforcing stimulus is contingent upon the expres-
shown that the neuromodulator released by the rein- sion of a designated behavior. The probability of ex-
forcing stimulus, which is believed to be serotonin, ac- pression of this behavior will then be altered. In other
tivates the enzyme adenylyl cyclase in the sensory words, the animal learns to associate the behavior
neuron and thereby increases the synthesis of the sec- with the contingent reinforcement and modifies its
ond messenger cAMP, which activates the cAMP- behavior accordingly.
APLYSIA: Classical Conditioning and Operant Conditioning 35

Behavioral Studies Figure 2

Feeding behavior in Aplysia has been used to gain
insights into the modification of a behavioral re-
sponse by reinforcement. Aplysia feed by protracting
a toothed structure called the radula into contact with
seaweed. The radula grasps seaweed by closing and
retracting. This sequence results in the ingestion of
the seaweed. Inedible objects can be rejected if the
radula protracts while closed (grasping the object)
and then opens as it retracts, releasing the object.
Thus, the timing of radula closure determines which
behavior will be elicited.
Feeding behavior in Aplysia can be modified by
pairing feeding with an aversive stimulus. In the pres-
ence of food wrapped in a tough plastic net, Aplysia
bite and attempt to swallow the food. However, netted
food cannot be swallowed, and it is rejected. The in-
ability to consume food appears to be an aversive
stimulus that modifies feeding behavior, since trained
animals do not attempt to bite netted food (Susswein,
Schwarz, and Feldman, 1986).
Feeding behavior can also be operantly condi-
tioned with an appetitive stimulus (Brembs et al., Model of operant conditioning of feeding in Aplysia. The
2002). Animals receiving positive reinforcement that cellular network that mediates feeding behavior is represented
by the elements in circles. Motor activity comprising two basic
is contingent on biting will learn to bite more than an-
feeding patterns is depicted above. A. At first, the radula
imals receiving either reinforcement that is not con-
protraction-generating element (Prot.) is active, followed by the
tingent on their behavior or no reinforcement at all. radula retraction element (Ret.). In the nave state, neuron B51
Neural Mechanisms of Appetitive Operant has a low probability for recruitment and thus does not take
part in the feeding motor program. Radula closure occurs
Conditioning in Aplysia
during the protraction phase. Consequently, the pattern elicited
The feeding system of Aplysia has many advan- is a rejection. B. Neuron B51 now has a higher probability for
tages. For example, much of the cellular circuitry con- recruitment following contingent reinforcement. B51 is now
trolling feeding behavior has been identified. Thus, active during the motor program, leading to radula closure
it is possible for researchers to study neurons with occurring primarily during the retraction phase. Thus, the
known behavioral significance. One of these neurons, pattern elicited will now be an ingestion.
denoted as B51, is implicated in the expression of in-
gestive behavior. B51 is active predominately during
the retraction phase. Furthermore, when B51 is re- patterns being produced. The contingent reinforce-
cruited into a pattern, it recruits radula closure motor ment also results in the modulation of the membrane
neurons (see Figure 2). properties of neuron B51. The input resistance in-
creases and a smaller stimulus is required to elicit
An in vitro analogue of operant conditioning has
electrical activity. Thus, the cell is more likely to be
been developed using only the isolated buccal gan-
active in the future. This change in the likelihood of
glia, which is responsible for generating the motor
B51 activation contributes to the conditioned in-
programs involved in feeding (Nargeot, Baxter, and
crease in ingestionlike patterns. Furthermore, these
Byrne, 1999a). The ganglion expresses motor pat-
results can be replicated when induced electrical ac-
terns that are analogous to feeding behavior. These
tivity in B51 is substituted as the analogue of the be-
motor patterns can either be ingestionlike or rejec-
havior, instead of an ingestionlike motor pattern
tionlike and the type that will be expressed is not pre-
(Nargeot, Baxter, and Byrne, 1999b).
dictable. In the analogue of operant conditioning,
motor patterns corresponding to ingestion are selec- The analogue can be further reduced by remov-
tively reinforced by contingently shocking the eso- ing neuron B51 from the ganglia and placing it in cul-
phageal nerve. The esophageal nerve contains dop- ture (Lorenzetti, Baxter, and Byrne, 2000; Brembs et
aminergic processes and is believed to be part of the al., 2002). This single, isolated neuron can be condi-
pathway mediating food reward. The conditioning tioned by contingently reinforcing induced electrical
results in an increase in the likelihood of ingestionlike activity (the analogue of behavior) with a direct and
36 APLYSIA: Classical Conditioning and Operant Conditioning

temporally discrete application of dopamine (the an- common both within any one species and between dif-
alogue of reinforcement). After conditioning, the ferent species? What is the relationship between the
membrane properties of B51 are again modulated initial induction of neuronal change (acquisition of
such that the cell is more likely to be active in the fu- learning) and the maintenance of the associative
ture. Such a highly reduced preparation is a promis- change? What are the relationships among different
ing candidate to study the mechanisms of dopamine- forms of learning, such as sensitization, classical con-
mediated reward and the conditioned expression of ditioning, and operant conditioning?
behavior at the level of the intracellular signaling cas-
One of the important findings to emerge from re- INSTRUMENTAL; GLUTAMATE RECEPTORS AND
cent studies on invertebrates is their capacity to ex- THEIR CHARACTERIZATION; INVERTEBRATE
hibit various forms of associative learning. Of particu-
lar significance is the finding that at least some
mollusks, such as Limax, exhibit higher-order features
of classical conditioning, such as second-order condi-
tioning and blocking. Contextual conditioning, con- Bibliography
ditioned discrimination learning, and contingency ef- Abrams, T. W., and Kandel, E. R. (1988). Is contiguity detection in
classical conditioning a system or cellular property? Learning
fects have been described in Aplysia (Colwill, Absher, in Aplysia suggests a possible site. Trends in Neurosciences 11,
and Roberts, 1988; Hawkins, Carew, and Kandel, 128135.
1986). Such higher-order features can be viewed in a Antonov, I., Antonova, I., Kandel, E. R., and Hawkins, R. D. (2001).
cognitive context, and raise the interesting possibility The contribution of activity-dependent synaptic plasticity to
that other complex behavioral phenomena will be classical conditioning in Aplysia. Journal of Neuroscience 21,
identified as the capabilities of these animals are in- Bao, J. X., Kandel, E. R., and Hawkins, R. D. (1998). Involvement
vestigated further. of presynaptic and postsynaptic mechanisms in a cellular ana-
log of classical conditioning at Aplysia sensory-motor neuron
The possibility of relating cellular changes to synapses in isolated cell culture. Journal of Neuroscience 18,
complex behavior in invertebrates is encouraged by 458466.
the progress that has already been made in examin- Brembs, B., Lorenzetti, F. D., Reyes, F. D., Baxter, D. A., and
ing the neural mechanisms of simple forms of nonas- Byrne J. H. (2002). Operant reward learning in Aplysia: Neu-
sociative and associative learning. The results of these ronal correlates and mechanisms. Science 296, 1,7061,709.
Carew, T. J., Hawkins, R. D., and Kandel, E. R. (1983). Differential
analyses of Aplysia have shown that: 1. learning in- classical conditioning of a defensive withdrawal reflex in Aply-
volves changes in existing neural circuitry (one does sia californica. Science 219, 397400.
not need the growth of new synapses and the forma- Carew, T. J., Walters, E. T., and Kandel, E. R. (1981). Classical con-
tion of new circuits for learning and short-term mem- ditioning in a simple withdrawal reflex in Aplysia californica.
ory to occur); 2. learning involves the activation of Journal of Neuroscience 1, 1,4261,437.
Colwill, R. M., Absher, R. A., and Roberts, M. L. (1988). Context-
second messenger systems; 3. the second messengers US learning in Aplysia californica. Journal of Neuroscience 8,
affect multiple subcellular processes to alter the re- 4,4344,439.
sponsiveness of the neuron (at least one locus for the Hawkins, R. D., Abrams, T. W., Carew, T. J., and Kandel, E. R.
storage and readout of memory is the alteration of (1983). A cellular mechanism of classical conditioning in Aply-
specific membrane currents); and 4. long-term mem- sia: Activity-dependent amplification of presynaptic facilita-
tion. Science 219, 400405.
ory requires new protein synthesis, whereas short- Hawkins, R. D., Carew, T. J., and Kandel, E. R. (1986). Effects of
term memory does not. interstimulus interval and contingency on classical condition-
ing of the Aplysia siphon withdrawal reflex. Journal of Neuro-
While researchers have made considerable prog-
science 6, 1,0951,701.
ress in the analysis of simple forms of learning in Aply- Lechner, H. A., Baxter, D. A., and Byrne, J. H. (2000). Classical
sia, other invertebrates, and vertebrate model sys- conditioning of feeding in Aplysia: I. Behavioral analysis. Jour-
tems, there is still no complete mechanistic analysis nal of Neuroscience 20, 3,3693,376.
available for any single example of simple learning. Lorenzetti, F. D., Baxter, D. A., and Byrne, J. H. (2000). Contin-
gent reinforcement with dopamine modifies the properties of
Many of the technical obstacles are being overcome,
an individual neuron in Aplysia. Society for Neuroscience Abstracts
however, and it is likely that the analyses of several ex- 26, 1,524.
amples of learning will reach completion. Murphy, G. G., and Glanzman, D. L. (1997). Mediation of classical
conditioning in Aplysia californica by long-term potentiation
For the near future, major questions to be an- of sensorimotor synapses. Science 278, 467471.
swered include the following: To what extent are Nargeot, R., Baxter, D. A., and Byrne J. H. (1999a). In vitro analog
mechanisms for classical and operant conditioning of operant conditioning in Aplysia: I. Contingent reinforce-
APLYSIA: Development of Processes Underlying Learning 37

ment modifies the functional dynamics of an identified neu- important insights into the final phenotypic expres-
ron. Journal of Neuroscience 19, 2,2472,260. sion of learning in the adult.
(1999b). In vitro analog of operant conditioning in Aplysia:
II. Modifications of the functional dynamics of an identified
neuron contribute to motor pattern selection. Journal of
Neuroscience 19, 2,2612,272.
The Development of Aplysia
Ocorr, K. A., Walters, E. T., and Byrne, J. H. (1985). Associative The life cycle of Aplysia can be divided into five
conditioning analog selectively increases cAMP levels of tail phases:
sensory neurons in Aplysia. Proceedings of the National Academy
of Sciences of the United States of America 82, 2,5482,552. 1. an embryonic phase (lasting about ten days, from
Schwartz, J. H., Bernier, L., Castellucci, V. F., Polazzolo, M., Sai- fertilization to hatching)
toh, T., Stapleton, A., and Kandel, E. R. (1983). What molecu-
lar steps determine the time course of the memory for short- 2. a planktonic larval phase (lasting about thirty-five
term sensitization in Aplysia? Cold Spring Harbor Symposium on days)
Quantitative Biology 48, 811819.
Susswein, A. J., Schwarz, M., and Feldman, E. (1986). Learned
3. a metamorphic phase (lasting only two to three
changes of feeding behavior in Aplysia in response to edible days)
and inedible foods. Journal of Neuroscience 6, 1,5131,527. 4. a juvenile phase (lasting at least ninety days)
Walters, E. T., and Byrne, J. H. (1983). Associative conditioning of
single sensory neurons suggests a cellular mechanism for 5. the adult phase, defined as the onset of reproduc-
learning. Science 219, 405408. tive maturity.
John H. Byrne These five phases can be further divided into thirteen
Revised by Fred D. Lorenzetti and John H. Byrne discrete stages, each defined by a specific set of mor-
phological criteria (Kriegstein, 1977). In the analysis
of learning that will be described here, most work
has focused on the juvenile phase of development
DEVELOPMENT OF PROCESSES (stages nine through twelve), since it is during this
UNDERLYING LEARNING time that many of the behavioral systems of interest
In the 1980s exciting progress was made in under-
standing a variety of developmental processes, rang-
ing from principles governing the birth, differentia- Forms of Learning and Developmental
tion, and migration of nerve cells to the mechanisms Timetables
underlying the functional assembly of complex neural
circuits. In addition to the intrinsic interest in devel- In adult Aplysia the siphon withdrawal reflex ex-
opment as a fundamental field of inquiry, the analysis hibits both nonassociative and associative forms of
of development has a secondary gain: By affording learning. The developmental analysis thus far carried
the experimental opportunity of examining early- out in Aplysia has focused primarily on nonassociative
emerging processes in functional isolation from later- learning in that reflex. The three most common
emerging ones, development can serve as a powerful forms of nonassociative learning are habituation, dis-
analytic tool with which to dissect and examine specif- habituation, and sensitization. Habituation refers to
ic behavioral, cellular, and molecular processes as a decrease in response magnitude occurring as a
they are expressed and integrated during ontogeny. function of repeated stimulation to a single site; dis-
habituation describes the facilitation of a habituated
A developmental strategy such as that described response by the presentation of a strong or novel
above has been useful in furthering the analysis of stimulus, usually to another site; sensitization refers
learning and memory in the marine mollusk Aplysia, to the facilitation of nondecremented responses by a
a preparation that has proved to be quite powerful for similar strong or novel stimulus. Using a behavioral
cellular and molecular studies of several forms of preparation that permitted quantification of siphon
learning. Specifically, using this strategy, it has been withdrawal throughout juvenile development, Rankin
possible to identify and dissociate multiple compo- and Carew (1987, 1988) found that these three forms
nents of nonassociative learning on both behavioral of nonassociative learning emerged according to dif-
and cellular levels. This type of analysis has also re- ferent developmental timetables. Habituation of si-
vealed previously unappreciated behavioral and cel- phon withdrawal was present very early (in stage nine)
lular processes in Aplysia. Moreover, the developmen- and progressively matured across all juvenile stages
tal dissociation of different components of learning in in terms of its interstimulus interval (ISI) function: In
juvenile Aplysia prompted a similar analysis in adult young animals, extremely short ISIs were necessary to
animals, where the same clearly dissociable compo- produce habituation, whereas in older animals, pro-
nents of learning were identified. Thus an analysis of gressively longer ISIs could be used to produce habit-
the developmental assembly of learning has provided uation. Dishabituation (produced by tail shock)
38 APLYSIA: Development of Processes Underlying Learning

Figure 1

Summary of developmental timetables for different forms of learning and their respective cellular analogues. Behavioral analysis:
Habituation is present as early as has been examined, in stage 9; dishabituation emerges soon after habituation, in stage 10;
sensitization does not emerge until sixty days after dishabituation, in mid to late stage 12. Behavioral inhibition has been measured as
early as stage 11 (indicated by shading) but may emerge even earlier. Cellular analysis: The cellular analogue of habituation
(homosynaptic decrement of EPSPs in neuron R2) is present in stage 9; the analogue of dishabituation (facilitation of decremented
EPSPs in R2) emerges in stage 10; the analogue of sensitization (facilitation of nondecremented EPSPs in R2) emerges between early
and mid stage 12. Inhibition of nondecremented EPSPs in R2 has been detected as early as early stage 12 (indicated by shading) but,
as with behavioral inhibition, may emerge even earlier. Thus there is a close developmental parallel between the emergence of each
behavioral form of learning (as well as behavioral inhibition) and its respective cellular analogue.

emerged soon after habituation, in a distinct and later The observation that dishabituation and sensiti-
stage (stage ten). However, sensitization (also pro- zation can be developmentally dissociated raises im-
duced by tail shock) did not emerge until surprisingly portant theoretical questions for a complete explana-
late in juvenile development (stage twelve), at least tion of nonassociative learning. For example, until
sixty days after the emergence of dishabituation (see recently a commonly held view was that nonassocia-
Figure 1). tive learning could be accounted for by a dual process
APLYSIA: Development of Processes Underlying Learning 39

theory involving two opposing processes: a single results illustrate two important points. First, the cellu-
decrementing process that gives rise to habituation, lar analogue of each form of learning emerges in
and a single facilitatory process that underlies both close temporal register with its respective behavioral
dishabituation and sensitization (Carew, Castellucci, form (see Figure 1). Second, just as dishabituation
and Kandel, 1971; Groves and Thompson, 1970). A and sensitization can be developmentally dissociated
key prediction of this view is that dishabituation and on a behavioral level, so can their cellular analogues
sensitization should always occur together. However, be developmentally dissociated.
the developmental dissociation of these processes, to-
gether with recent behavioral and cellular evidence in
adult Aplysia, suggests that a dual-process view is inad- A Novel Inhibitory Process
equate to account for nonassociative learning in Aply- When the effects of sensitization training (i.e., the
sia. effects of tail shock on nondecremented reflex re-
The emergence of sensitization in stage twelve is sponses) in different developmental stages were ex-
not confined to the siphon withdrawal reflex in Aply- amined, an unexpected effect of tail shock was discov-
sia. Stopfer and Carew (1988) examined another re- ered: Prior to the emergence of sensitization in stage
sponse system, escape locomotion, and found that twelve, tail shock had an inhibitory effect on reflex re-
sensitization in that system also emerges in stage sponsiveness (Rankin and Carew, 1988). The proper-
twelve. Thus sensitization is expressed in two differ- ties of this inhibitory process in juvenile Aplysia have
ent systems, one a graded reflex and the other a cen- been studied by Rankin and Carew (1989), who found
trally programmed cyclical behavior, at the same time that tail shock-induced inhibition of siphon withdraw-
in development. This raises the interesting hypothe- al can be detected in two ways: 1. by reduction of re-
sis that one or more developmental signals may flex responsiveness; and 2. by the apparent competi-
switch on the general process of sensitization in stage tion of the inhibitory process with the facilitatory
twelve, not only in individual response systems but in process of dishabituation. Specifically, they found
the whole animal. that as levels of tail shock were increased, progressive-
ly more inhibition resulted and, concomitantly, pro-
gressively less dishabituation was produced, suggest-
Cellular Analogues of Learning and ing the hypothesis that the tail shock-induced
Behavioral Learning inhibition could significantly retard the expression of
dishabituation in early developmental stages. Finally,
The developmental separation of different learn- as the process of sensitization matured, there was a
ing processes described above provides the opportu- clear transition from the inhibitory effect of tail shock
nity to examine the unique contribution of specific to reflex facilitation between early and late stage
cellular mechanisms to each form of learning. An im- twelve.
portant step in such a cellular analysis is to show that
the cellular analogue of each form of learning can be In parallel with the behavioral reflection of inhi-
identified in the central nervous system of juvenile bition described above, Nolen and Carew (1988)
Aplysia and that these analogues exhibit a develop- identified a clear cellular analogue of this inhibitory
mental time course parallel to the behavioral expres- process. Specifically, they found that prior to the
sion of the learning. The identified motor neuron can emergence of the cellular analogue of sensitization in
serve as a reliable cellular monitor of plasticity in the mid to late stage twelve, activation of the pathway
afferent input to the siphon-withdrawal reflex. from the tail produced significant inhibition of non-
decremented synaptic responses in neuron R2 (see
The developmental emergence of the cellular an- Figure 1). As with the behavior, there was a clear tran-
alogue of habituation (synaptic decrement) and of sition from inhibition to facilitation in mid to late
dishabituation (facilitation of decremented synaptic stage twelve.
potentials) was first examined by Rayport and Camar-
do (1984). They found that synaptic decrement could The inhibitory process first identified in juvenile
be observed in neuron R2 as early as stage nine, and Aplysia has received considerable attention in the
that facilitation of depressed synaptic potentials adult. Several laboratories have observed behavioral
emerged in stage ten. Nolen and Carew (1988) then tail shock-induced inhibition of the siphon withdraw-
examined the emergence of the cellular analogue of al reflex (Krontiris-Litowitz, Erikson, and Walters,
sensitization (facilitation of nondecremented synaptic 1987; Mackey et al., 1987; Marcus et al., 1988), and
potentials) in R2. They found that the analogue of important progress has been made in studying the
sensitization emerged between early and middle cellular mechanisms underlying the inhibitory pro-
stage twelve, many weeks after the emergence of the cess. For example, Mackey et al. (1987) found that tail
analogue of dishabituation. Taken collectively, these shock produced presynaptic inhibition of the trans-
40 APLYSIA: Development of Processes Underlying Learning

mission from siphon sensory neurons. Wright, Mar- stronger stimuli. Moreover, the stimulus intensity
cus, and Carew (1991) found that polysynaptic input that was most effective in producing dishabituation
to the siphon motor neurons plays an important role produced no sensitization, and the intensity that was
in mediating tail shock-induced inhibition, and most effective in producing sensitization produced no
Wright and Carew (1990) found that a single identi- significant dishabituation. Thus, as in juvenile Aplysia
fied inhibitory interneuron in the abdominal gangli- (Rankin and Carew, 1989), the processes of dishabi-
on, cell L16, can account for most, if not all, of the in- tuation, sensitization, and inhibition can be behavior-
hibition of siphon withdrawal following tail shock. ally dissociated in adult animals.
Finally, Fitzgerald and Carew (1991) found that sero-
tonin, a known facilitatory neuromodulator in Aplysia, The behavioral observations described above
can also mimic the inhibitory effects of tail shock. It raise important questions about the cellular processes
will be of considerable interest to study the develop- underlying the dissociation of dishabituation and sen-
ment of these inhibitory mechanisms and examine sitization. One possibility is that these two forms of
the way in which they are integrated with facilitatory learning reflect different underlying cellular mecha-
forms of behavioral plasticity. nisms. Alternatively, the same or related mechanisms
may be involved in both forms of learning, and the
dissociation observed could be due to differential in-
Behavioral Dissociation of Dishabituation, teraction of the inhibitory process with dishabituation
Sensitization, and Inhibition in Adults and sensitization. Behavioral results alone cannot dis-
tinguish between these possibilities. However, prog-
The developmental studies described above show ress has been made in elucidating the cellular mecha-
that dishabituation, sensitization, and a novel inhibi- nisms underlying dishabituation and sensitization
tory process, as well as their respective cellular ana- (Hochner et al., 1986) as well as inhibition (Bellardet-
logues, can each be dissociated in juvenile animals. It ti, Kandel, and Siegelbaum, 1987; Mackey et al.,
is possible, however, that these processes, although 1987; Wright and Carew, 1990; Wright, Marcus, and
separable during ontogeny, are not distinct in the Carew, 1991). Thus it will be important to determine
final adult phenotype. Thus an important question the degree to which these different cellular processes
arose as to whether the same forms of behavioral plas- can account for the behavioral dissociations that are
ticity could be identified and separated in adult ani- observed in both developing and adult Aplysia.
mals. Marcus et al. (1988) addressed this issue by ex-
amining, in adult Aplysia, the effects of a wide range
of tail-shock intensities, at several times after tail Conclusion
shock, on both habituated and nonhabituated siphon
A developmental analysis in Aplysia has shown
withdrawal responses. They found that dishabituation
that different forms of learning, as well as their cellu-
and sensitization could be clearly dissociated in adult
lar analogues at central synapses, emerge according
animals in two ways.
to very different developmental timetables. These
First was time of onset. When tested soon (ninety studies have allowed the dissociation of four behavior-
seconds) after tail shock, dishabituation was evident al processes (see Figure 1): two decrementing (habitu-
at a variety of stimulus intensities, whereas, in this ation and inhibition) and two facilitatory (dishabitua-
early test, sensitization was not exhibited at any stimu- tion and sensitization). Whether these dissociations
lus intensity. In fact, examining nondecremented re- are produced by different facilitatory mechanisms, by
sponses revealed that tail shock produced inhibition differential interactions of inhibition with decre-
of reflex amplitude. Although no sensitization was ev- mented and nondecremented responses, or by some
ident in the ninety-second test, in subsequent tests combination of these alternatives, results suggest that
(twenty to thirty minutes after tail shock) significant a dual-process view of nonassociative learning, which
sensitization was observed. Thus, dishabituation has postulates a single decremental and a single incre-
an early onset (within ninety seconds), whereas sensi- mental process, requires revision, and that a multiple-
tization has a very delayed onset (twenty to thirty min- process view, which includes the possibility of inhibi-
utes) after tail shock. Juvenile Aplysia also exhibit de- tory as well as facilitatory interactions, is necessary to
layed-onset sensitization that emerges in early stage account adequately for the mechanisms underlying
twelve, at least thirty days after the emergence of dis- nonassociative learning.
The developmental studies discussed in this brief
Second was stimulus intensity. When a range of review have focused only on nonassociative learning
stimulus intensities to the tail was examined, maximal in Aplysia, and only on short-term learning, which is
dishabituation was produced by relatively weak stimu- retained for a relatively brief time (minutes to hours).
li, whereas maximal sensitization was produced by However, Aplysia is also capable of exhibiting a variety
APLYSIA: Molecular Basis of Long-Term Sensitization 41

of forms of associative learning. Moreover, in addi- Marcus, E. A., Nolen, T. G., Rankin, C. H., and Carew, T. J. (1988).
tion to short-term forms, both nonassociative and as- Behavioral dissociation of dishabituation, sensitization, and
inhibition in Aplysia. Science 241, 210213.
sociative learning in Aplysia can exist in long-term Nolen, T. G., and Carew, T. J. (1988). A cellular analog of sensitiza-
forms lasting days to weeks. It will be of interest to ex- tion emerges at the same time in development as behavioral
amine the development of these additional processes sensitization in Aplysia. Journal of Neuroscience 8, 212222.
in order to gain insights into theoretically important Rankin, C. H., and Carew, T. J. (1987). Development of learning
questions such as the relationships between nonasso- and memory in Aplysia: II. Habituation and dishabituation.
Journal of Neuroscience 7, 133144.
ciative and associative learning and between short- (1988). Dishabituation and sensitization emerge as separate
term and long-term memory. As a step in this direc- processes during development in Aplysia. Journal of Neuro-
tion, Carew, Wright, and McCance (1989) have estab- science 8, 197211.
lished that long-term memory for sensitization (1989). Developmental analysis in Aplysia reveals inhibitory
emerges in exactly the same stage as short-term mem- as well as facilitatory effects of tail shock. Behavior and Neuro-
science 103, 334344.
ory, stage twelve. This observation lends support to Rayport, S. G., and Camardo, J. S. (1984). Differential emergence
the notion that short- and long-term memory may be of cellular mechanisms mediating habituation and sensitiza-
mechanistically interrelated in Aplysia (Golet et al., tion in the developing Aplysia nervous system. Journal of
1986). By analyzing the development of these diverse Neuroscience 4, 2,5282,532.
processes at synaptic, biophysical, and molecular le- Stopfer, M., and Carew, T. J. (1988). Development of sensitization
of escape locomotion in Aplysia. Journal of Neuroscience 8, 223
vels, it may be possible to gain unique insights into 230.
the substrates underlying learning and memory by Wright, W. G., and Carew, T. J. (1990). Contributions of interneu-
examining their developmental assembly. rons to tail-shock induced inhibition of the siphon withdrawal
reflex in Aplysia. Society for Neuroscience Abstracts 16, 20.
Wright, W. G., Marcus, E. A., and Carew, T. J. (1991). A cellular
See also: APLYSIA: CLASSICAL CONDITIONING AND analysis of inhibition in the siphon withdrawal reflex of Aply-
OPERANT CONDITIONING; APLYSIA: MOLECULAR sia. Journal of Neuroscience 11, 2,4982,509.
Thomas J. Carew

Bellardetti, F., Kandel, E. R., and Siegelbaum, S. (1987). Neuronal
inhibition by the peptide FMRFamide involves opening of S
K+ channels. Nature 325, 153156.
Carew, T. J., Castellucci, V. F., and Kandel, E. R. (1971). Analysis SENSITIZATION
of dishabituation and sensitization of the gill withdrawal re-
flex in Aplysia. International Journal of Neuroscience 2, 7998. Sensitization is a simple form of nonassociative learn-
Carew, T. J., Wright, W. G., and McCance, E. F. (1989). Develop- ing that involves the enhancement of the response to
ment of long-term memory in Aplysia: Long-term sensitiza- a weak stimulus that occurs after the presentation of
tion is present when short-term sensitization first emerges. So- a strong or noxious stimulus. Sensitization usually oc-
ciety for Neuroscience Abstracts 15, 1,285. curs in two forms that differ in their duration and un-
Fitzgerald, K., and Carew, T. J. (1991). Serotonin mimics tail shock
in producing transient inhibition in the siphon withdrawal re- derlying mechanisms. Short-term sensitization lasts
flex of Aplysia. Journal of Neuroscience 11, 2,5102,518. seconds to minutes and involves the modification of
Golet, P., Castellucci, V. F., Schacher, S., and Kandel, E. R. (1986). neuronal membrane properties and synaptic efficacy,
The long and short of long-term memorya molecular often through the alteration of the phosphorylation
framework. Nature 322, 419422. state of existing proteins. Long-term sensitization
Groves, P. M., and Thompson, R. F. (1970). Habituation: A dual
process theory. Psychological Review 77, 419450. lasts from days to weeks, depending on the training
Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986). Ad- protocol. Unlike the short-term version, long-term
ditional component in the cellular mechanism of presynaptic sensitization requires synthesis of new macromo-
facilitation contributes to behavioral dishabituation in Aplysia. leculesthe inhibition of either gene transcription
Proceedings of the National Academy of Sciences of the United States into mRNA or translation of mRNA into protein
of America 83, 8,7948,798.
Kriegstein, A. R. (1977). Stages in post-hatching development of blocks long-term sensitization. In its most persistent
Aplysia californica. Journal of Experimental Zoology 199, 275288. form, long-term sensitization involves morphological
Krontiris-Litowitz, J. K., Erikson, M. T., and Walters, E. T. (1987). changes and neuronal growth.
Central suppression of defensive reflexes in Aplysia by noxious
stimulation and by factors released from body wall. Society for The marine mollusk Aplysia has proved a useful
Neuroscience Abstracts 13, 815. model for gaining insights into the underlying neural
Mackey, S. L., Glanzman, D. L., Small, S. A., Dyke, A. M., Kandel, and molecular mechanisms of long-term sensitiza-
E. R., and Hawkins, R. D. (1987). Tail shock produces inhibi- tion. Aplysia has a simple nervous system with large,
tion as well as sensitization of the siphon-withdrawal reflex of individually identifiable neurons that are accessible
Aplysia: Possible behavioral role for presynaptic inhibition me-
diated by the peptide Phe-Met-Arg-Phe-NH2. Proceedings of for detailed anatomical, biophysical, biochemical,
the National Academy of Sciences of the United States of America 84, and molecular studies. Researchers have identified
8,7308,734. many of the neural circuits involved in sensitization
42 APLYSIA: Molecular Basis of Long-Term Sensitization

of several reflexive withdrawal behaviors and have duced preparations derived from sensitized and con-
characterized individual neurons within these cir- trol animals have shown that changes in the animals
cuits. They have also identified specific locations of behavior correlate with changes in the properties of
learning-related modifications in these circuits and both the presynaptic sensory neuron and the postsyn-
have analyzed the underlying cellular mechanisms. aptic motor neuron. Although these are unlikely to be
the only sites of plasticity associated with sensitization
in the nervous system, changes in the sensory and
The Siphon Withdrawal Reflexes of Aplysia motor neurons are likely to be important because
Sensitization studies in Aplysia have focused chief- they occur in the basic circuit of the reflex arc.
ly on the withdrawal reflexes of the siphon gill and
the tail siphon. The gill, the animals respiratory
organ, is in the mantle cavity. The siphon is a fleshy, Modulation of the Presynaptic Sensory
tubelike extension of the body wall protruding from Neuron
the mantle cavity. A stimulus delivered to the siphon The first correlate of long-term sensitization to
elicits withdrawal of both the siphon and the gill: This be described was a decrease in sensory-neuron-
is the siphon-gill withdrawal reflex. Eliciting the tail- membrane potassium current (Scholz and Byrne,
siphon withdrawal reflex requires the delivery of a 1987). Researchers had found that modulation of the
stimulus to the tail, causing withdrawal of both the tail same current for short-term sensitization was sensi-
and the siphon. In the wild, sensitization might be in- tive to serotonin (5-hydroxytryptamine, 5-HT) and
duced by an unsuccessful attack by a predator (e. g., cyclic adenosine monophosphate (cAMP) treatments.
pinching or scratching by a crab). In the laboratory, The decrease in membrane current after long-term
electrical shocks to the body wall are the most com- sensitization training leads to an increase in the num-
mon means of sensitizing the animal because it is easi- ber of action potentials elicited for a fixed amount of
er to deliver the same stimulus repeatedly. Long-term injected current. Thus, sensory neurons fire more ac-
sensitization requires prolonged training (multiple tion potentials in response to a given stimulus after
shocks delivered over an extended intervale.g., one sensitization (Cleary, Lee, and Byrne 1998).
to two hours). One training session can sensitize the
animal for several days, whereas multiple training ses- Another correlate of long-term sensitization is an
sions repeated over several days can sensitize the ani- increase in synaptic efficacy at the synapses between
mal for weeks. sensory neurons and motor neurons (Cleary, Lee, and
Byrne 1998; Frost, Castellucci, Hawkins, and Kandel,
The afferent signal is carried by mechanosensory
1985). A major component of this synaptic modula-
neurons. About twenty-four sensory neurons inner-
tion appears to be due to an increase transmitter re-
vate the siphon, and another 200 sensory neurons in-
lease from the presynaptic sensory neurons (Dale,
nervate the body wall (a subset of only about twenty
Schacher, and Kandel, 1988), although the nature of
of these innervate the tail). These sensory neurons
this increase is still poorly understood. The prime tar-
make excitatory synapses onto motor neurons that
gets of of modulation are as follows: broadening ac-
cause contractions of muscles in the gill, siphon, and
tion potential, increased numbers of release-ready
tail. Sensory neuron membrane properties and pre-
synaptic vesicles, and increased efficiency of the syn-
synaptic release machinery are the principal sites of
aptic-release machinery. It is not yet known whether
plasticity during sensitization. Sensory neurons also
other synapses made by sensory neurons (e.g., onto
synapse with a variety of interneurons. Some inter-
interneurons) are also enhanced. In addition, post-
neurons convey sensory information from one body
synaptic mechanisms may also contribute to in-
region to the circuits that control other regions.
creased synaptic efficacy (Cleary, Lee, and Byrne
Other interneurons secrete modulatory transmitters
1998; see below).
such as serotonin that modify the biophysical proper-
ties of sensory and motor neurons and the character- More extensive training (e.g., four days) pro-
istics of neurotransmitter release from the sensory duces more persistent sensitization. After four days of
neuron (i.e., heterosynaptic facilitation, discussed in training, anatomical studies have revealed remodel-
more detail below). ing of sensory neuron release sites and the growth of
additional axonal and dendritic branches as well as
new synaptic contacts that may contribute to in-
Cellular Correlates of Long-Term creased synaptic efficacy (Bailey and Chen, 1983;
Sensitization Wainwright, Zhang, Byrne, and Cleary, 2002). The
One advantage of Aplysia for studies on long-term recruitment of this growth requires several days of
learning is that the nervous system, or portions of it, training rather than a single day (Wainwright, Zhang,
can be removed for studies in vitro. Studies in re- Byrne, and Cleary, 2002). The down regulation of the
APLYSIA: Molecular Basis of Long-Term Sensitization 43

Aplysia neuronal cell adhesion molecule, ApCAM, 1999; Zhang et al., 1997). Furthermore, a scavenger
plays an important role in this growth-associated molecule specific for transforming growth factor
plasticity (Abel and Kandel, 1998). Sensory neurons blocks long-term facilitation induced by 5-HT treat-
reduce their synthesis of new ApCAM protein and in- ment. These data suggest that the Aplysia homologue
ternalize ApCAM already present at their plasma of the growth factor may be necessary for the induc-
membranes. The signal for internalization of ApCAM tion or expression of long-term plasticity by acting
is phosphorylation by extracellular signal-regulated downstream of 5-HT. Transforming growth factor
kinase (ERK, discussed below). The decrease in activates two protein kinases (PKC and ERK, see
ApCAM in sensory-neuron axonal membranes may below) in Aplysia sensory neurons (Chin et al., 2002;
allow the formation of new branches and additional Farr, Mathews, Zhu, and Ambron 1999)
neurotransmitter release sites.

Second Messengers and Protein Kinase

Modulation of the Postsynaptic Neuron Cascades
Changes in the properties of the postsynaptic In the early 1980s, researchers found that sensiti-
neuron may also contribute to the increase in synaptic zation training and 5-HT each increased cytoplasmic
efficacy underlying long-term sensitization. After levels of the intracellular second messenger, cAMP
training, there are changes in two motor-neuron (Bernier, Castellucci, Kandel, and Schwartz, 1982;
membrane properties: an increase in the resting- Ocorr, Tabata, and Byrne, 1986). Regulation of this
membrane potential and a decrease in the spike second messenger is critical for long-term memory in
threshold (Cleary et al, 1998). In addition, motor fruit flies, bees, and rodents. In Aplysia, injection of
neurons show a protein-synthesis-dependent in- cAMP into sensory neurons produces many of the
creased responsiveness to glutamate twenty-four long-term changes previously correlated with sensiti-
hours after 5-HT treatment (Trudeau and Castelluc- zation training, including decreased potassium con-
ci, 1995). This result suggests an increase in the num- ductance, increased membrane excitability, and neu-
ber of postsynaptic receptors that accompanies the in- rite growth (OLeary, Byrne, and Cleary, 1995;
crease in presynaptic transmitter release. However, Schacher et al., 1988, 1993; Scholz and Byrne, 1988).
blocking this increase in postsynaptic responsiveness Indeed, cAMP appears to be sufficient for the induc-
does not block long-term facilitation. tion of several of the major forms of long-term neuro-
nal plasticity associated with sensitization.
Heterosynaptic Modulation The key effector of cAMP for long-term neuronal
What is the modulatory signal that produces plasticity is the cAMP-dependent protein kinase
these presynaptic and postsynaptic changes? It seems (PKA). Inhibition of PKA during the induction and
that the sensory neurons that convey information early stages of consolidation of long-term memory
about the sensitizing stimulus into the central nervous blocks synaptic facilitation one day after 5-HT treat-
system synapse onto one or more interneurons. These ment (Abel and Kandel, 1998), and injection of the
interneurons, in turn, release a modulatory transmit- active catalytic subunit of PKA can induce long-term
ter to produce the changes described above. A grow- facilitation in the absence of upstream signaling
ing body of evidence indicates that the transmitter (Chain et al., 1999). Among the substrates of PKA that
used by these interneurons is serotonin (5-HT) are critical for the induction of long-term plasticity
(Byrne and Kandel, 1996). Indeed, 5-HT mimics the are several transcription factors that regulate gene
many effects of sensitization training: Prolonged or expression. The principal target of PKA is the cAMP/
multiple applications of 5-HT are necessary to pro- Ca2+-responsive element binding protein (CREB,
duce long-term changes in vitro. As an analog of sen- discussed below), a transcription factor that plays im-
sitization, 5-HT use has proved invaluable in studies portant roles in many forms of long-term memory. In
seeking to delineate the cellular mechanisms that un- addition, PKA activity persists at least twenty-four
derlie long-term sensitization. hours after sensitization training or treatment with
5-HT (Chain et al., 1999; Mller and Carew, 1998).
In addition to 5-HT, other transmitter molecules
It is possible, then, that persistent kinase activity is
may induce long-term sensitization. In the late 1990s,
also important for long-term changes.
researchers found evidence that a member of the
transforming growth factor family was involved in Although the activation of the cAMP/PKA cascade
long-term sensitization (Zhang et al., 1997). Indeed, appears to be sufficient to produce long-term plastici-
treatment of Aplysia neurons with human transform- ty associated with sensitization, other kinases play im-
ing growth factor results in long-term facilitation portant roles in the induction of sensitization. One of
and increased sensory-neuron excitability (Chin et al., the kinases is protein kinase C (PKC). Sensitization
44 APLYSIA: Molecular Basis of Long-Term Sensitization

training or extended 5-HT treatment results in a pro- Immediate Early Genes

longed activation of PKC that lasts about two hours CREB-dependent transcription leads to the ex-
after training or treatment. During this period PKC
pression of immediate early genes. Researchers have
contributes to the regulation of translation and the
identified two of these genes. One is a key component
activation of a transcription factor (C/EBP, see below)
of the ubiquitin-proteosome pathway for the degra-
that regulates the expression of late genes necessary
dation of proteins, Aplysia ubiquitin C-terminal hy-
for long-term sensitization.
drolase (ApUCH), and the other is another transcrip-
Another kinase cascade that plays an essential tion factor, CCAAT/enhancer binding protein (C/
role in the induction of long-term facilitation is the EBP). ApUCH is an enzyme that associates with the
extracellular signal-regulated kinase (ERK), a mem- proteosome and increases its proteolytic activity.
ber of the mitogen-activated protein kinase family. ApUCH is an immediate early gene involved in the
ERK can be activated by several intracellular cas- induction of long-term sensitization (Abel and Kan-
cades. Researchers have yet to pinpoint the one that del, 1998; Alberini, 1999). CREB activation increases
leads to its activation during sensitization training or the expression of ApUCH. An important substrate of
by 5-HT treatment. Yet in Aplysia sensory neurons the proteosome in sensory neurons is the regulatory
cAMP can mediate both the activation of ERK and its subunit of PKA. Increased expression of ApUCH re-
translocation to the nucleus. The cAMP cascade ap- sults in the proteosome-mediated degradation of a
pears to be important for activation of ERK during subset of the PKA regulatory subunits (those binding
the early stages of long-term sensitization induction. cAMP). As the ratio of the regulatory to catalytic sub-
ERK can also be activated by PKC, a cascade that may units decreases, PKA activity becomes independent of
be important during later stages of induction. cAMP and is prolonged. Thus, CREB activation leads
The protein synthesis required for long-term sen- to degradation of the PKA regulatory subunit and
sitization occurs during a critical period that begins thereby prolongs the activity of the PKA catalytic sub-
at the onset of training and continues until training unit, which is now independent of cAMP.
ends (Castellucci, Blumenfeld, Goelet, and Kandel, C/EBP, the CCAAT/enhancer binding protein, is
1989; Levenson et al., 2000). This period corre- another transcription factor. In Aplysia sensory neu-
sponds to the time when CREB-dependent transcrip- rons, C/EBP is an immediate early gene expressed in
tion occurs. response to CREB activation (Alberini, 1999). C/EBP
There is evidence that a second round of protein can form both homodimers with other C/EBP pro-
synthesis occurs during the formation of long-term teins or heterodimers with other transcription factors
sensitization. Growth of sensory neurons in response to activate transcription of late genes. In order to acti-
to cAMP injections is blocked by protein-synthesis in- vate transcription of its target genes, C/EBP must be
hibitor up to seven hours after injection (OLeary, phosphorylated by ERK. Although the synthesis of a
Byrne, and Cleary, 1995). This second round of pro- number of proteins are known to be regulated by sen-
tein synthesis may correspond to the gene expression sitization, researchers have yet to identify C/EBP tar-
that stabilizes long-term sensitization. get genes in Aplysia.

Although there are other sites of neural plasticity,

CREB and Memory the detailed study of the modifications that occur in
sensory neurons and their synapses with motor neu-
As noted above, long-term memory differs from
rons (i. e., the fundamental reflex arc) during long-
short-term memory chiefly in requiring protein syn-
term sensitization have led to an understanding of the
thesis. During the induction of long-term memory,
basic mechanisms that underlie this simple form of
activated PKA translocates from the sensory neuron
learning. Research using the relatively simple ner-
cytoplasm into the nucleus, where it phosphorylates
vous system of Aplysia has provided insights into
the transcriptional activator CREB1 (Abel and Kan-
memory formation at the cellular, molecular, and bio-
del, 1998). CREB1 activates transcription by binding
to the DNA of certain genes that contain cAMP/Ca2+ physical levels. These basic mechanisms appear to
responsive element (CRE) sequences in their regula- form the platform upon which more complex forms
tory domains. Another form of CREB, CREB2, re- of learning (i. e., classical and operant conditioning)
presses transcription of these genes. CREB2 is a sub- develop.
strate of ERK, which also translocates into the nucleus
during the induction phase. Thus, expression of See also: MORPHOLOGICAL BASIS OF LEARNING AND
CRE-containing genes requires both activation by MEMORY: INVERTEBRATES; PROTEIN SYNTHESIS

Abel, T., and Kandel, E. R. (1998). Positive and negative regulatory
mechanisms that mediate long-term memory storage. Brain
Research Reviews 26, 360378.
Alberini, C. (1999). Genes to remember. Journal of Experimental Bi-
ology 202, 2,8872,891.
Bailey, C. H., and Chen, M. (1983). Morphological basis of long-
term habituation and sensitization in Aplysia. Science 220, 91
Bernier, L., Castellucci, V. F., Kandel, E. R,. and Schwartz, J. H.
(1982). Facilitatory transmitter causes a selective and pro-
longed increase in adenosine 3,5-monophosphate in sensory
neurons mediating the gill and siphon withdrawal reflex in
Aplysia. Journal of Neuroscience 2, 1,6821,691.
Byrne, J. H., and Kandel, E. R. (1996). Presynaptic facilitation re-
visited: State and time dependence. Journal of Neuroscience 16,
Castellucci, V. F., Blumenfeld, H., Goelet, P., and Kandel, E. R.
(1989). Inhibitor of protein synthesis blocks long-term behav-
ioral sensitization in the isolated gill-withdrawal reflex of Aply-
sia. Journal of Neurobiology 20, 19.
Chain, D. G., Casadio, A., Schacher, S., Hegde, A. N., Valbrun, M.,
Yamamoto, N., Goldberg, A. L., Bartsch, D., Kandel, E. R.,
and Schwartz, J. H. (1999). Mechanisms for generating the
autonomous cAMP-dependent protein kinase required for
long-term facilitation in Aplysia. Neuron 22, 147156.
Chin, J., Angers, A., Cleary, L. J., Eskin, A., and Byrne, J. H.
(1999). TGF-1 in Aplysia: Role of long-term changes in the
excitability of sensory neurons and distribution of TR-II-like
immunoreactivity. Learning and Memory 6, 317330.
(2002). TGF-1 alters synapsin distribution and modulates Aristotle (The Library of Congress)
synaptic depression in Aplysia. Journal of Neuroscience 22,
Cleary, L. J., Lee, W. L., and Byrne, J. H. (1998). Cellular corre-
changes produced by neurotransmitters in Aplysia sensory
lates of long-term sensitization in Aplysia. Journal of Neuro-
neurons. Neuron 10, 1,0791,088.
science 18, 5,9885,998.
Scholz, K. P., and Byrne, J. H. (1987). Long-term sensitization in
Dale, N., Schacher, S., and Kandel, E. R. (1988). Long-term facili-
Aplysia: biophysical correlates in tail sensory neurons. Science
tation in Aplysia involves increase in transmitter release. Sci- 235, 685687.
ence 239, 282285. (1988). Intracellular injection of cAMP induces a long-term
Farr, M., Mathews, J., Zhu, D. F., and Ambron, R. T. (1999). In- reduction of neuronal K+ currents. Science 240, 1,6641,666.
flammation causes a long-term hyperexcitability in the noci- Trudeau, L.-E., and Castellucci, V.F. (1995). Postsynaptic modifi-
ceptive sensory neurons of Aplysia. Learning & Memory 6, 331 cations in long-term facilitation in Aplysia: Upregulation of
340. excitatory amino acid receptors. Journal of Neuroscience 15,
Frost, W. N., Castellucci, V. F., Hawkins, R. D., and Kandel, E. R. 1,2751,284.
(1985). Monosynaptic connections made by the sensory neu- Wainwright, M. L., Zhang, H., Byrne, J. H., and Cleary, L. J.
rons of the gill- and siphon-withdrawal reflex in Aplysia partic- (2002). Localized neuronal outgrowth induced by long-term
ipate in the storage of long-term memory for sensitization. sensitization training in Aplysia. Journal of Neuroscience 22,
Proceedings of the National Academy of Sciences of the United States 4,1324,141.
of America 82, 8,2668,269. Zhang, F., Endo, S., Cleary, L. J., Eskin, A., and Byrne, J. H.
Levenson, J., Endo, S., Kategaya, L. S., Fernandez, R. I., Brabham, (1997). Role of transforming growth factor- in long-term
D. G., Chin, J., Byrne, J. H., and Eskin, A. (2000). Long-term synaptic facilitation in Aplysia. Science 275, 1,3181,320.
regulation of neuronal high-affinity glutamate and glutamine
uptake in Aplysia. Proceedings of the National Academy of Sciences Vincent Castellucci
of the United States of America 97, 12,85812,863. Revised by Gregg A. Phares and John H. Byrne
Ocorr, K. A., Tabata, A. M., and Byrne, J. H. (1986). Stimuli that
produce sensitization lead to an elevation of cyclic AMP levels
in tail sensory neurons of Aplysia. Brain Research 371, 190192.
OLeary, F. A., Byrne, J. H., and Cleary, L. J. (1995). Long-term
structural remodeling in Aplysia sensory neurons requires de
novo protein synthesis during a critical time period. Journal
ARISTOTLE (384322 B.C.E.)
of Neuroscience 15, 3,5193,525. Aristotle was born in northern Greece, in the town of
Schacher, S., Castellucci, V. F., and Kandel, E. R. (1988). cAMP
Stagira, in 384 B.C.E. At seventeen, he went to Athens
evokes long-term facilitation in Aplysia sensory neurons that
requires new protein synthesis. Science 240, 1,6671,669. and became a student in Platos Academy, where he
Schacher, S., Kandel, E. R., and Montarolo, P. (1993). cAMP and remained for twenty years. Although greatly influ-
arachidonic acid simulate long-term structural and functional enced by Plato and by the pre-Socratic philosophers,

especially Empedocles, Aristotle was a highly original ception of a state of affairs can be stored, 2. how it can
thinker and a disciple of no one. In 347 B.C.E. he left be brought to mind later, and 3. how it happens that,
Athens and traveled extensively in Asia Minor, be- when it is brought to mind, the relation between the
coming tutor to Alexander the Great in 342 B.C.E. representation and the original state of affairs, now
Seven years later he returned to Athens and began his absent, is such that the first is a memory of the second
own school, the Lyceum. After the death of Alexander and is known to be such. In contemporary dress, these
the Great in 323 B.C.E., he left Athens, and he died are the problems of information storage, information
the following year in Khalks, a few miles north of retrieval, and the general problem of how representa-
Athens. tions represent.
In his main work on memory, De memoria et re- Aristotle tries to explain information storage by
miniscentia, Aristotle tries to dissect out the central appeal to the analogy of imprinting soft wax with a
phenomena to be explained, and suggests mechani- seal. He reasons that sense perception is somehow
cal explanations of a general sort to account for them. like a picture and that it is the perception picture that
In his scientific works, Aristotle typically seeks the re- stamps its likeness to create a memory. Apparently
ality behind the appearances, and he expects that the the perception is stamped on the soul (Aristotle has
reality may be different from what it seems. This is es- a physical, not a supernatural, conception of the soul),
pecially forward-looking in the case of mental phe- or at any rate, it is stamped on some sort of physical
nomena, where subsequent thinkers, such as Ren stuff that can be in causal interaction with it and can
Descartes (in the seventeenth century) and Zeno take on some of its properties. This helps address the
Vendler (in the late twentieth century), insist that representation problem. The imprint (memory rep-
mental reality must be exactly as it seems. Aristotles resentation) resembles, physically, the perception
collection of memory phenomena displays some sys- (perceptual representation), which in turn resembles,
tematicity, and with characteristic insight, he lights on physically, that of which it is a perception. So by tran-
several basically correct classifications. Nevertheless, sitivity of resemblance, there is a correlation between
to modern eyes some of his collection is a bit of a jum- stored representation and original state of affairs. Ar-
ble, and the mechanical explanations tendered are so istotles conclusion that there must be a resemblance
implausible that they must have been no more than was taken as axiomatic by most subsequent thinkers,
helpful metaphors to him. and they searched for the parameters of physical re-
semblance. Research since the 1970s, especially in
Aristotles relentlessly naturalistic perspective, computer science and neuroscience, has revealed that
however, gives him a decidedly modern stamp. That representation does not require resemblance in any
is, he sought physical rather than supernatural or straightforward sense, a radical departure from earli-
spiritual explanations for memory phenomena, and er theories.
he well knew the importance of observations even
though his own were occasionally mere assumptions. In asking how representations represent, Aristot-
(For example, he thought women had fewer teeth le identified a truly fundamental problem. Still only
than men.) In the absence of a developed biology, ex- partially solved, it remains a central problem, though
perimental psychology, or neuroscience, he could it is now addressed within the framework of modern
hardly be expected either to envisage explanations in psychology, philosophy, neuroscience, and computer
terms of neuronal connectivity or to know how to pen- science.
etrate learning phenomena at the behavioral level. Understanding the importance of broad systema-
ticity in a theory, Aristotle tests a theorys strength by
seeing how much can be encompassed within its
Observations and Explanations ambit. Thus he claims that the stuff that receives the
In commenting upon memory and learning phe- imprint may have varying degrees of imprintability.
nomena, Aristotles fundamental distinction is be- Explanations are then forthcoming for ones poor
tween recalling information to mind and storing in- recollection of early childhood and for declining
formation, or, as he puts it, between remembering, memory in the elderly: In very young children the
which is the reinstatement in consciousness of some- stuff is too much like running water to take the im-
thing that was there before (451b6), and memory, print; in older humans, the stuff hardens and no lon-
the existence, potentially, in the mind (452a10), of ger is very impressible. Extending this idea further,
an earlier perception or conception. In modern par- Aristotle thinks a related explanation will apply to his
lance, this is the distinction between remembering in observation that those who are too quick and those
the occurrent sense and remembering in the stored who are too slow also have poor memories. Exactly
or dispositional sense. The central problems, in Aris- what phenomenon he is addressing here is unclear,
totles view, are to explain three things: 1. how a per- and this may be one of those inexplicable Aristotelian

observations that need a much broadened base of with the result that the animal can recognize different
data. individuals as belonging to the same category. In hu-
The representation problem, Aristotle notices, mans this means, for example, that a pine tree, a yew,
has a further dimension. When an image from memo- and an olive tree may all be recognized as similar de-
ry comes to mind, how do we know that it is a memo- spite differences in shape, size, and color. He says that
ry, rather than a thought or image without relation to the soul is so constituted that the universal tree can
bygone events? That is, how does the occurrent pre- be developed from the stored perceptions of individ-
sentation carry the information that it is a memory? ually distinct items. A slug, on the other hand, lacks
His answer has two parts. First, sometimes we do get the capacity to generalize across individuals because
confused, and we think a presentation is a memory it lacks the capacity to store information.
when it is not (false memory); and sometimes we have In Aristotles view, storing information provides
a memory presentation but are unaware that it is a the similarity substructure that underpins both scien-
memory. So the system is imperfect. Second, when tific categorization and the skilled knowledge dis-
the system does work, it is because for animals with played by craftsmen who can make many different
memory, the organ whereby they perceive time is clay pots or ships captains who can sail under many
also that whereby they remember (449b30). The different conditions. In modern guise, his idea is that
idea here is that when perceptions are stored as mem- generalization to items that are relevantly similar but
ories, they are also somehow indexed as to time, so incidentally different, both perceptually and behav-
that the imprint bears not only the perceptions shape iorally, requires information storage. Additionally, he
but also its whenness. regards this capacity as enabling experience, the rea-
Retrieval appears to require something like an son being that experience requires understanding,
image or an iconic presentation that resembles the which in turn requires categorization of perceptions.
original perception. The mechanism of retrieval Consequently, animals such as humans have genuine
should, one surmises, have to do with something tak- experience; animals such as slugs do not (Posterior An-
ing up the stored imprint and re-presenting it, but in alytics, 99b36;l00a5).
fact Aristotle says nothing of this. Instead he discusses
the phenomenon of association, noting that events Conclusion
experienced together are often remembered togeth- Any inclination to feel smug about Aristotles
er. He explains associated recollections by saying that shortcomings should be tempered by noting that
the movement of a perception causes the move- even current classifications of learning phenomena
ment of the memory. He sees, therefore, that part of are controversial and tentative, and experimental
the theory of storage will include the relations be- psychologists are sometimes chided for doing little
tween associated memories, but he neither provides more than codifying common sense. Nor, of course,
an account of those storage relations nor elaborates should Aristotle himself be blamed for the slavish
on how information is retrieved by the movements adoption of his every word by uncritical monks in the
(451b1530). Middle Ages. Aristotle the scientist-philosopher was
In Historia animalium Aristotle suggests that hu- anything but dogmatic. Twentieth-century physical
mans and animals differ in that humans alone can re- explanationsalthough not mechanical, but electri-
member something at will (488b25), though he also cal and biochemicalsit well with his abiding natural-
notes in De memoria et reminiscentia that recollections ism.
can occur without effort. Indeed, he observes that For a long period in the history of thought, Aris-
melancholics often have obsessive memories, try totles views on nearly everything were taken as au-
though they might to repress them. In the physicalist thoritative. His Metaphysics probably had the greatest
spirit, he conjectures that melancholics have more impact; however, the work on memory was not espe-
moisture around their sense perception center, which cially influential.
is easily set in motion, thus explaining the memorys
being presented again and again despite ones will. Bibliography
Aristotle (1941). De anima, trans. J. A. Smith. In Richard Mc-
Aristotle believed that animals differ in whether Keon, ed., The basic works of Aristotle. New York: Random
they have the capacity to store their perceptions; ani- House.
mals with the capacity to do so have genuine knowl- (1941). De memoria et reminiscentia, trans. by J. I. Beare. In
edge of their world, whereas animals lacking the ca- Richard McKeon, ed., The basic works of Aristotle. New York:
Random House.
pacity merely respond to their current perceptions on
(1941). Historia animalium. In Richard McKeon, ed., The
the basis of their innate dispositions. The advantage basic works of Aristotle. New York: Random House.
of storing perceptions is that the stored items may (1975). Aristotles Posterior analytics, trans. J. Barnes. Oxford:
come to have systematic relations among themselves, Clarendon Press.

Beare, J. I. (1906). Greek theories of elementary cognition. Oxford: tention left over for other inputs or other tasks. Ordi-
Clarendon Press. nary usage also seems to imply that limitations of at-
Churchland, P. S. (2002). Brain-wise: Studies in neurophilosophy.
Cambridge, MA: MIT Press.
tention are a key factor restricting ones ability not
Descartes, R. (1649; reprint 1968). Les passions de lme. English only to perceive many stimuli at once, but also to per-
translation in E. S. Haldane and G. R. T. Ross, trans. (1911), form two tasks at the same time. These commonplace
The philosophical works of Descartes. Cambridge, UK: Cam- notions from folk psychology may or may not accu-
bridge University Press. rately describe humans information-processing ma-
Edel, A. (1982). Aristotle and his philosophy. Chapel Hill: University
of North Carolina Press. chinery.
Kahn, C. H. (1966). Sensation and consciousness in Aristotles psy-
Late-twentieth-century research argues that the
chology. Archiv fr Geschichte der Philosophie 48, 4381.
Ross, W. D. (1923; reprint 1959). Aristotle. Cleveland: Meridian phenomena of attention are not as unitary as com-
Books. mon sense might suggest (Pashler, 1998). This is
(1955). Aristotles Parva naturalia. Oxford: Oxford Univer- scarcely surprising given the complexities of the
sity Press. bodys underlying neural machinery and of the tasks
Sorabji, R. (1972). Aristotle on memory. London: Trinity Press.
to which the human brain is adapted. Evidence for
Patricia Smith Churchland distinct attentional mechanisms is seen most clearly
Georgios Anagnostopoulos in relation to divided attention: the performance limita-
Revised by Patricia Smith Churchland tions that arise when a person attempts to process
more than one stimulus at a time. On the one hand,
there appear to be a set of processing limitations asso-
ciated with perceptual analysis of inputs in different
sense modalities. If a person is confronted with differ-
ent sensory inputs at the same time, it is more difficult
It seems to be a tenet of ordinary common sense that to perceive them when they arrive through the same
people remember what they attend to and forget what sense modality rather than through different modali-
they do not. Not surprisingly, researchers have noted ties. This was first demonstrated by A.Treisman and
the very close relationship between attention and A. Davies (1973), who showed that people were better
memory for a very long time, and some empirical evi- able to monitor animal names when some of the
dence for the linkage was offered as far back as the words to be monitored were presented visually and
late nineteenth century (Smith, 1895). However, it others auditorily (as compared to both in the same
was only during the twentieth century, with the ad- modality). These modality-restricted perceptual ca-
vent of cognitive psychology and its relatively rich pacity limitations are wholly governed by the volun-
array of methods for studying human information tary direction of selective attention: It is only the stim-
processing over fine time scales, that it became possi- uli that are attended to that compete for limited
ble to begin to analyze this connection in more detail. resources, not all the stimuli that may be impinging
To do so, researchers have used taxonomies of me- on the senses.
morial and attentional processes that emerge from
laboratory studies in each of these areas. In addition to the perceptual processing limita-
tions tied to particular input modalities, research
points to a separate set of limitations that become evi-
Forms of Attention dent only when a person tries to perform multiple
The term attention as used in everyday language tasks at the same time. These central attentional limi-
is a diffuse and global term that alludes to both selec- tations have their locus in the more cognitive stages
tivity and capacity limitations. The potential for selectiv- of processing, especially the planning of actions and
ity is evident in the fact that of the great multitude of the retrieval of information in memory. When two
stimuli impinging on the sense organs at any one in- tasks each require mental operations of these types,
stant, human beings are usually vividly aware of only the processing in the two tasks is normally subject to
a fairly small subset. Capacity limitations are evident queuing: selection of a response in one task must be
whenever people try to attend to more than one completed before selection of the response in the
stream of inputs, particularly if comprehension or re- other task can commence (Pashler and Johnston,
sponse is required: for example, trying to listen to the 1998). Interestingly, these limitations seem quite in-
radio at the same time as one reads a newspaper. Ca- different to what modality the information arrives in.
sual usage seems to imply that attention refers to a If a person must make a speeded response to a tone,
single process or substance that accounts for both se- for example, this will delay his or her ability to make
lectivity and capacity limitations. Thus, people speak a rapid response to a concurrently presented color
of focusing attention on one sensory input or one patch. The central interference is also independent of
task, with the result that one does not have enough at- response modalities; for example, if one response is

vocal and one manual, the interference is still ob- information is normally retained in echoic memory
served. for one to two seconds. Sensory memory systems,
When one seeks to understand the relation of at- while impressive in capacity, do not retain informa-
tention to memory, it is fruitful to inquire both about tion long enough to be useful for most purposes. If
the role of modality-specific perceptual attention the information is not transferred to short-term and/
mechanisms and the role of central attentional mech- or long-term memory, it is lost.
anisms. Does information get into sensory memory even
when a person attempts to ignore it at the time it is
presented? For auditory sensory memory, the answer
Forms of Memory
is evidently yes. This is seen, for example, in early ob-
Memory, too, appears to be composed of a num- servations by Broadbent and others that when people
ber of relatively separate functional systems. Several shadow spoken input to one ear, they can, if inter-
key distinctions have been proposed since the mid- rupted, abruptly switch over and (relying on echoic
1960s. The most important and well validated of memory) recall the last bit of information presented
these is the distinction between three broad memory to the other ear. Daily life offers many examples of
systems that hold information over different time the same phenomenon. Although there is little re-
scales (and differ in certain other properties). This search that directly addresses the question of atten-
analysis, sometimes called the modal model, postu- tional involvement in iconic memory storage, the data
lates separate sensory memory, short-term memory that do exist suggest that even unattended informa-
(STM), and long-term memory (LTM) systems. The tion is briefly represented in this literal memory stor-
model arose out of the pioneering work of N. Waugh age.
and D. A. Norman (1965) and M. Glanzer and A. R.
Cunitz (1966). Sensory memory systems seem to pro-
duce something akin to a brief literal persistence of Attention and Short-Term Memory
a stimulus for a very short time beyond its actual pre-
Primary or short-term memory maintains infor-
sentation. This probably reflects continued firing of
mation in an active state, but only so long as it is at
neurons in sensory pathways after the offset of a stim-
least periodically rehearsed or refreshed. While some
ulus. Short-term (or working) memory refers to a set
theorists originally proposed a single unitary STM, it
of very limited-capacity storage mechanisms. It is
has become clear that there are a number of separate
often assumed to reflect an actively refreshed neural
and independent short-term memory systems. One
representation in sensory, perceptual, and motor-
system (sometimes termed the articulatory buffer and
control areas, perhaps maintained by a reverberative
familiar to everyone who has ever remembered a
process involving frontal brain structures. Long-term
phone number for a half minute or so) holds speech-
memory, on the other hand, refers to the relatively
like representations of verbal material. Another kind
more permanent set of memory traces that is almost
of STM system holds visual information in a schemat-
surely encoded by changes in synaptic weights. In ad-
ic form. There are hints that other independent sys-
dition to the three-part demarcation, many investiga-
tems may also exist, holding, for example, nonvisual
tors propose a distinction within the realm of long-
spatial representations, acoustic imagery, semantic
term memory, distinguishing between so-called ex-
representations, and tactile or kinesthetic patterns.
plicit memory (underlying conscious recollections, as
The clearest evidence that different STM systems are
in recall or recognition tasks) and implicit memory
indeed independent systems comes from experi-
(e.g., changes caused by exposure to a stimulus that
ments showing that people can hold onto more than
modulate later processing without causing conscious
one type of information at the same time. For exam-
ple, people can retain spoken digits and visual mate-
rials with little mutual interference. Further clinching
Attention and Sensory Memory the case is the finding that some patients, after suffer-
ing brain damage, have lost one form of STM without
Sensory memory systems can potentially retain a any loss of the others (Basso, Spinnler, Vallar, and
large amount of detailed sensory information about Zanobio, 1982).
an input for an extremely short period. For visual in-
puts, the sensory memory is called iconic memory. Icon- Many researchers refer to working memory to
ic memory seems to hold onto inputs for only about indicate that information in STM may be actively ma-
100 to 400 milliseconds, depending on physical prop- nipulated or transformed during the time it is re-
erties of the stimulus and the visual input that follows tained. Whatever purpose information is put to and
it. Sensory memory for auditory sensations is usually whatever term may be used for it, the most notable
called echoic memory, and most investigators agree that characteristic of STM storage is limited capacity. For

spoken material, a very limited number of words (or getting it out (retrieval). Devoting perceptual atten-
perhaps more critically, syllables) can be retained. tion to a stimulus is by no means sufficient for LTM
For visual patterns, even a four-by-four checkerboard storage (whereas it may be sufficient for STM). A per-
grid is enough to overload visual short-term memory. son can see one hundred words exposed one at a time
Access time for STM is quite good, however. For ex- in rapid succession (at a rate of, say, four per second),
ample, it appears that the image of a letter can be successfully detecting every word in the list that is an
transferred into visual short-term memory in a matter animal name, and at the end he or she will often re-
of a few hundred milliseconds, and the rate for speech member scarcely anything except animal names. Ob-
is probably not much different. What is the connec- viously the person attended to all the words in order
tion between attention and STM storage? Attending to determine their semantic category, but the process
to a stimulus is clearly a necessary condition for stor- left no permanent residue.
ing that stimulus in short-term memory. Less clear is
What does produce memory storage, then? Many
whether attending to a stimulus is a sufficient condi-
studies have concluded that the key factor is elabora-
tion. The critical experiment would involve asking
tion: active processing that uncovers connections be-
whether someone can retain one set of information
tween a to-be-stored item and other information in
(A) in STM and then attend to additional information
long-term store (Craik and Lockhart, 1972). By con-
arriving in the same modality (B) without having B
trast, mere intention or desire to remember seems
overwrite A. Using visual patterns, W. A. Phillips and
relatively unimportant.
F. M. Christie (1977) concluded that this may be pos-
sible but that the issue requires further study. As for Whereas central attentional mechanisms may
storage of spoken information, attending to irrele- play only a fairly minor role in the encoding of infor-
vant speech often compromises short-term memory mation into STM, for LTM they appear critical. Near-
quite considerably, suggesting that mere attention to ly every study that has looked at peoples ability to
a new input overwrites the existing contents of STM. form long-term memory traces while performing a
concurrent task has found a substantial impairment.
Does the storage and retention of information in This occurs even when there is no discernible similari-
STM require central attentional mechanisms dis- ty between the stimuli to be remembered and those
cussed above? For articulatory STM, the involvement used in the concurrent task: for example, remember-
appears to be intermittent rather than continuous ing odors while playing a computer game (Perkins
(perhaps arising in some initial consolidation in STM, and Cook, 1990).
and then later in scheduling periodic rehearsals). A
person can retain a phone number, for example, and In the late 1990s there was a lively controversy
perform another brief demanding task involving un- about the role of central attentional mechanisms in
related materials without losing the phone number. retrieval of information from long-term memory. On
However, if the demanding task is initiated immedi- the one hand, when people are given unlimited time
ately after storing some spoken material, memory for to retrieve materials, a concurrent task often fails to
this material may suffer (Naveh-Benjamin and Jo- much dent the number of items they can produce
nides, 1984), and the same can happen when people (Craik, Govoni, Naveh-Benjamin, and Anderson,
undertake a continuously demanding task for some 1996). On the other hand, a demanding concurrent
sustained period. task markedly reduces the rate at which information
can be retrieved from LTM (Hicks and Marsh, 2000).
The picture that emerges, then, is of short-term Furthermore, when an unrelated speeded-choice re-
memory systems closely tied to perceptual input sys- action-time task is performed concurrently with a
tems (and corresponding perceptual attention ma- paired-associate retrieval task, there is evidence that
chinery). Central attentional limitations may have the memory retrieval is completely delayed by central
some involvement, but it is relatively indirect or inter- attention to the speeded task (Carrier and Pashler,
mittent. 1995).

Attention and Long-Term Memory Attention and Implicit Memory

To store information in long-term memory, one The idea of a separate implicit memory system
need not do anything active to maintain it. Nonethe- grew out of data showing amnesics may have a spared
less, recollecting a memory tends to strengthen its ability to form certain kinds of memories: those for
long-term memory representation, often dramatical- which explicit recollection is not required (or at least
ly so; conversely, memories are probably subject to a subset of such memories). A number of studies re-
erosion with the mere passage of time. What seems to ported that formation of implicit memories also does
be most difficult is getting information into LTM, and not require limited-capacity attentional resources.

For example, M. E. Smith and M. Oscar-Berman Rajaram, S., Srinivas, K., and Travers, S. (2001). The effects of at-
(1990) observed that a concurrent task at the time of tention on perceptual implicit memory. Memory & Cognition
29, 920930.
word encoding did not reduce the priming effect Smith, M. E., and Oscar-Berman, M. (1990). Repetition priming
found for repeated words as tested in a later lexical of words and pseudowords in divided attention and in amne-
decision task. However, more recent studies tend to sia. Journal of Experimental Psychology: Learning, Memory, and
find that there are considerable costs when more de- Cognition 16, 1,0331,042.
manding secondary tasks are performed at the time Smith, W. G. (1895). The relation of attention to memory. Mind 4,
of encoding (Mulligan and Hornstein, 2000; Ra- Treisman, A., and Davies, A. (1973). Dividing attention to ear and
jaram, Srinivas, and Travers, 2001). Researchers have eye. In S. Kornblum, ed., Attention and performance IV. New
more to learn about this conflict. It is possible that ex- York: Academic Press.
plicit memory measures are simply more sensitive to Waugh, N., and Norman, D. A. (1965). Primary memory. Psycholog-
impaired storage produced by divided attention, but ical Review 72, 89104.
it is also possible that implicit memory storage reflects Harold Pashler
mechanisms that are independent of central atten-
tional capacity.



Basso, A., Spinnler, H., Vallar, G., and Zanobio, M. E. (1982). Left See: MODALITY EFFECTS
hemisphere damage and selective impairment of auditory
verbal short-term memory: A case study. Neuropsychologia 20,
Broadbent, D. E. (1957). Immediate memory and simultaneous
stimuli. Quarterly Journal of Experimental Psychology 9, 111. AUTOBIOGRAPHICAL MEMORY
Carrier, L. M., and Pashler, H. (1995). Attentional limits in memo-
Autobiographical memory is the psychological history
ry retrieval. Journal of Experimental Psychology: Learning, Memo-
ry, and Cognition 21, 1,3391,348. of the self. It consists of memories of personal experi-
Craik, F. I. M., Govoni, R., Naveh-Benjamin, M., and Anderson, encesepisodic memoriesand knowledge of the
N. D. (1996). The effects of divided attention on encoding self or autobiographical knowledge: for example,
and retrieval processes in human memory. Journal of Experi- schools we attended, people we had relationships
mental Psychology: General 125, 159180.
with, places we have lived, places we have worked, and
Craik, F. I. M., and Lockhart, R. S. (1972). Levels of processing:
A framework for memory research. Journal of Verbal Learning so on (Conway, 2001; Conway and Pleydell-Pearce,
and Verbal Behavior 11, 671684. 2000; and McAdams, 2001). It is critical for personal
Glanzer, M., and Cunitz, A. R. (1966). Two storage mechanisms in identity, forming the basis of the self and binding self-
free recall. Journal of Verbal Learning and Verbal Behavior 5, conceptions to reality. Psychiatric illnesses or brain
damage can disrupt the connections that bind self to
Hicks, J. L., and Marsh, R. L. (2000). Toward specifying the atten-
tional demands of recognition memory. Journal of Experimen- reality through memory, leading to a loss of personal
tal Psychology: Learning, Memory, and Cognition 26, 1,483 history and the attendant delusions, confabulations,
1,498. and false beliefs.
Mulligan, S., and Hornstein, S. (2000). Attention and perceptual
priming in the perceptual identification task. Journal of Experi-
mental Psychology: Learning, Memory, and Cognition 26, 626 The Nature of Autobiographical Memory
637. and Its Relation to Self
Naveh-Benjamin, M., and Jonides, J. (1984). Maintenance rehears-
al: A two-component analysis. Journal of Experimental Psycholo- Autobiographical knowledge encompasses far
gy: Learning, Memory, and Cognition 10, 369385. more than memory: It includes statements, proposi-
Pashler, H. E. (1998). The psychology of attention. Cambridge, MA: tions, declarations, and beliefs about the self, often
MIT Press.
Pashler, H., and Johnston, J. C. (1998). Attentional limitations in accompanied by generic and/or specific (mainly visu-
dual-task performance. In H. Pashler, ed., Attention. Hove, al) images of details of prior experience. Autobio-
East Sussex, UK: Psychology Press. graphical knowledge is distinct from sensory percep-
Perkins, J., and Cook, N. M. (1990). Recognition and recall of tual episodic memories, which represent details of
odours: The effects of suppressing visual and verbal encoding actual experience (Conway, 2001). In the formation
processes. British Journal of Psychology 81, 221226.
Phillips, W. A., and Christie, F. M. (1977). Interference with visual- of an autobiographical memory, autobiographical
ization. Quarterly Journal of Experimental Psychology 29, 637 knowledge becomes linked to episodic memories
650. (Conway and Pleydell-Pearce, 2000). Memory forma-

tion leads to recollective experience, a sense or feel- There have been many attempted theoretical ex-
ing of the self in the past (Tulving, 1985; Wheeler, planations of childhood amnesia (Pillemer and
Stuss, and Tulving, 1997), and attention turns inward White, 1989), but most founder on young childrens
to the autobiographical memory and, perhaps, to capacity for a wide range of specific episodic memo-
other episodic memories and autobiographical ries and detailed autobiographical knowledge (Fivush
knowledge. Of course, full autobiographical memory et al., 1996). Explanations that center on changes in
formation does not have to take place, and autobio- intellect, language, and emotional development fail
graphical knowledge and episodic memory can be simply because apparently normal autobiographical
processed independently. memories are typically accessible before the age of
Conway and Pleydell-Pearce (2000) apply the five and only after that seem to submerge in a general
term the working self to the control structure that mod- forgetfulness; it seems unlikely that an increase in
ulates this whole system of autobiographical memory general functioning would make unavailable already
formationthe dynamic combining of autobiograph- accessible memories. Hence there is no compelling
ical knowledge with episodic memories. The working explanation for this component of the lifespan-
self consists of an active goal hierarchy (only parts of retrieval curve, which remains a challenge to autobio-
which are consciously accessible), models or concep- graphical memory researchers.
tions of the self, and other forms of self-knowledge
that facilitate access to autobiographical knowledge The reminiscence bump has also attracted its
structures. New autobiographical knowledge and epi- share of theories (Rubin, Rahhal, and Poon, 1998).
sodic memories are formed (encoded) through the One is that this period endures in memory because it
working self, which also influences memory construc- is suffused with novel experiences. An alternative ex-
tion by controlling input to the knowledge base and planation holds that although only a small percentage
by evaluating output (activated autobiographical of experiences during the reminiscence bump are
knowledge). The working self may even exercise in- novel events, they survive because of their uniqueness
hibitory control over the knowledge base (Conway
in the formation of a persons life circumstances and
and Pleydell-Pearce, 2000).
interests (Fitzgerald, 1988); on this view, it is the high
accessibility of memories from this period that ac-
Autobiographical Memory Across the counts for their durability (Conway and Pleydell-
Lifespan Pearce, 2000). Perhaps many memories from the pe-
The working selfgoal hierarchy and self- riod of the reminiscence bump are of self-defining
conceptionsprobably first emerges in some coher- experiences (Fitzgerald, 1988; Singer and Salovey,
ent form as the infant develops, in its second year, the 1993) and have a powerful effect in binding the work-
capacity for objective and subjective self-awareness in ing self to a specific reality. But this period, like chlid-
the form of conceptions of I and me. (Howe and hood amnesia, has yet to find a definitive explanation
Courage, 1997). Children as young as thirty months for its relation to memory.
have detailed autobiographical memories (Fivush,
Hadden, and Reese, 1996), although these are not The recency component of the lifespan-retrieval
ususally accessible in adulthood. Undoubtedly the curve (see Figure 1) can be simply explained as a peri-
working self and its relation to autobiographical od of forgetting: recently encoded memories, whose
memory changes during childhood and perhaps sta- accessibility is retained for a longer interval, are sub-
bilises into an enduring form only in late adolescence ject to decay and/or interference and so become pro-
and early adulthood (Erikson, 1950). gressively less accessible. This is a familiar, often-
These periods of development of the self are re- observed pattern of retention. One might wonder,
flected in the lifespan-retrieval curve that arises from however, why such memories or salient experiences
older adults (about thirty-five years and older) free or should be forgotten in this way. If older adults are
cued recall of autobiographical memories (Franklin expressly instructed to recall autobiographical mem-
and Holding, 1977; Fitzgerald and Lawrence, 1984; ories from this period of forgetting, there are appar-
Rubin, Wetzler, and Nebes, 1986; Rubin, Rahhal, and ently plenty of available memories (Holmes and Con-
Poon, 1998). This technique plots the age of encod- way, 1999). Thus, this seems a matter not of
ing of memories; as shown in Figure 1, the lifespan- forgetting but rather of bias or preference in access
retrieval curve consists of three periods: childhood to memories. It may be that the recency portion of the
amnesia, (from birth to approximately five years of lifespan-retrieval curve reflects a lowering in the self-
age), reminiscence bump (from ten to thirty years), relevance of memories of recent experiences and
and the period of recency (from the present declining hence a lowering in their accessibility rather than
back to the period of the reminiscence bump). complete forgetting.

Figure 1

Autobiographical Memory and Personality erativity. In contrast, those participants without a

The working self increases the accessibility of prominent commitment story showed no such bias. In
a similar way work by Woike and her colleagues has
goal-related autobiographical knowledge. Markus
further established the connection between personali-
(1977) found preferential access to memories of expe-
ty and memory (Woike, 1995; Woike, Gershkovich,
riences congruent with central self-schema, those that
Piorkowski, and Polo, 1999). Woike and colleagues
are critical to someones sense of independence or de-
(1999) investigated groups of individuals classified as
pendence. McAdams (1982) identified individuals
agentic (concerned with personal power, achieve-
with a strong intimacy motivation or with a distinctive
ment, and independence) or as communion (con-
power motivation and found that the intimacy-
cerned with relationships, interdependence, and oth-
motivation group recalled peak experiences with a ers). Agentic types consistently recalled emotional
preponderance of intimacy themes compared to indi- memories of events that involved issues of agency
viduals who scored lower on this motivation, who (mastery, humiliation). In contrast, communal types
showed no memory bias. Similarly, the power- recalled emotional memories featuring others, often
motivation group recalled peak experiences with significant others, in acts of love and friendship.
strong themes of power and satisfaction. Subsequent- These and a range of findings from other studies
ly, McAdams, Diamond, de Aubin, and Mansfield, (McAdams, 2001) all show a that the dominant mo-
(1997) examined the influence of the Eriksonian no- tives or goals of the self make memories of goal-
tion of generativity (Erikson, 1963) on the life stories relevant experiences highly accessible.
of middle-aged adults. Generativity refers to nurtur-
ing and caring for those things, products, and people
that have the potential to outlast the self. Those indi- Autobiographical Memory in Distress
viduals who were judged high in generativity, who Brain injury can impair autobiographical memo-
had a commitment story, were found to recall a pre- ry in various ways (Conway and Fthenaki, 2000). Inju-
ponderance of events highly related to aspects of gen- ries to regions of the frontal lobes often lead to a

clouding or loss of detailed memories. In more ex- Fitzgerald, J. M., and Lawrence, R. (1984). Autobiographical mem-
treme cases patients may confabulate, constructing ory across the lifespan. Journal of Gerontology 39, 692698.
Fivush, R., Haden, C., and Reese, E. (1996). Remembering, re-
autobiographical knowledge into plausible but false
counting, and reminiscing: The development of autobio-
memories. Patients with damage to the temporal graphical memory in social context. In D.C. Rubin, ed., Re-
lobes and underlying structures in the limbic system, membering our past: Studies in autobiographiocal memory, pp. 341
especially the hippocampal formation, may lose the 359. Cambridge: Cambridge University Press.
ability to form new memories while retaining access Franklin, H. C., and Holding, D. H. (1977). Personal memories at
to at least some preinjury memories. Those with dam- different ages. Quarterly Journal of Experimental Psychology 29,
age to posterior regions of the brain, regions involved 527532.
Holmes, A., and Conway, M. A. (1999). Generation identity and the
in visual processing (occipital lobes) may lose the abil-
reminiscence bump: Memories for public and private events.
ity to generate visual images of the past and, because Journal of Adult Development 6, 2134.
of this, become amnesic. Their amnesia occurs be- Howe, M. L., and Courage, M. L. (1997). The emergence and early
cause the episodic content of autobiographical mem- development of autobiographical memory. Psychological Re-
ories is predominantly encoded in the form of visual view 104, 499523.
images. When the ability to generate visual images is Markus, H. (1977). Self-schemata and processing information
compromised or lost because of brain damage, then about the self. Journal of Personality and Social Psychology 35,
access to specific details of the past held in episodic
McAdams, D. P. (1982). Experiences of intimacy and power: Rela-
images is also lost. tionships between social motives and autobiographical memo-
In psychiatric illness a common occurrence is that ry. Journal of Personality and Social Psychology 42, 292302.
of a severe clouding of autobiographical memory, re- (2001). The psychology of life stories. Review of General Psy-
sulting in overgeneral memories. For instance, in chology 5, 100122.
McAdams, D. P., Diamond, A., de Aubin E., and Mansfield, E.
clinical depression patients recall many memories
(1997). Stories of commitment: The psychosocial construction
that lack detail and are much more schematic than of generative lives. Journal of Personality and Social Psychology
typical autobiographical memories. Thus, a patient 72, 678694.
asked to recall specific memories of his father could Rubin, D. C., Rahhal, T. A., Poon, L. W. (1998). Things learned in
only recall general events such as walks in the park early adulthood are remembered best. Memory & Cognition 26,
after Sunday lunch but was unable to generate a sin- 319.
gle specific memory of a single walk (Williams, 1996). Rubin, D. C., Wetzler, S. E., and Nebes, R. D. (1986). Autobio-
graphical memory across the adult lifespan. In D. C. Rubin,
Clouded, overgeneral memories have also been ob-
ed., Autobiographical memory. Cambridge, UK: Cambridge
served in schizophrenic patients and in patients suf- University Press.
fering from obsessional-compulsive disorder. One Singer, J. A., and Salovey, P. (1993). The remembered self. New York:
possibility is that the complex control processes that The Free Press.
modulate memory construction (working self) be- Tulving, E. (1985). Memory and consciousness. Canadian Psycholo-
come attenuated in psychiatric illnesses and so can no gist 26, 112.
longer form fully detailed memories. Wheeler, M. A., Stuss, D. T., and Tulving, E. (1997). Towards a the-
ory of episodic memory: The frontal lobes and autonoetic
consciousness. Psychological Bulletin 121, 351354.
Williams, J. M. G. (1996). Depression and the specificity of autobio-
Conway, M. A. (2001). Sensory perceptual episodic memory and its
graphical memory. In D. C. Rubin, ed., Remembering our past:
context: Autobiographical memory. Philosophical Transactions
Studies in autobiographical memory. Cambridge, UK: Cambridge
of the Royal Society of London 356, 1,2971,306.
University Press.
Conway, M.A., and Fthenaki, A. (2000). Disruption and loss of au-
Woike, B. (1995). Most-memorable experiences: Evidence for a
tobiographical memory. In L. S. Cermak, ed., Handbook of
link between implicit and explicit motives and social cognitive
neuropsychology: Memory and its disorders. Amsterdam: Elsevier.
Conway, M. A., and Pleydell-Pearce, C. W. (2000). The construc- processes in everyday life. Journal of Personality and Social Psy-
tion of autobiographical memories in the self memory system. chology 68, 1,0811,091.
Psychological Review 107, 261288. Woike, B., Gershkovich, I., Piorkowski, R., and Polo, M. (1999).
Erikson, E.H. (1950). Childhood and society. New York: W. W. Nor- The role of motives in the content and structure of autobio-
ton. graphical memory. Journal of Personality and Social Psychology
Fitzgerald, J. M. (1988). Vivid memories and the reminiscence 76, 600612.
phenomenon: The role of a self narrative. Human Development
31, 261273. Martin A. Conway
BARTLETT, FREDERIC (18861969) tion but also in the retrieval of memories. Another
interest concerned perception and memory perfor-
Frederic Charles Bartlett was born on October 20, mance in people of other cultures, including South
1886, in Stow on the Wold, Gloucestershire. He stud- Africa, a country he had visited. His discussions of
ied literature, logic, and philosophy before becoming conventionalization leaned heavily on evidence col-
a tutor at the University of Cambridge in 1909. At lected in other societies; in the late twentieth century
Cambridge, his interests turned to psychology, partly there was a revival of interest in Bartletts contribu-
through his acquaintance with James Ward; he was tions to cross-cultural psychology (Saito, 1999).
awarded a fellowship at St. Johns College in 1913
and obtained a first-class degree in moral sciences in In 1922 Myers left Cambridge to head the Na-
1914. Cambridge then was in the forefront of the tional Institute for Industrial Psychology, and Bartlett
movement to make experimental psychology a recog- was appointed reader and director of the Cambridge
nized branch of science in the British university sys- laboratory; two years later he also became editor of
tem; C. S. Myers (18731947), a lecturer in experi- the British Journal of Psychology, a position he held for
mental psychology there, not only had campaigned twenty-four years. In 1925 he wrote an important
for Cambridge to build a laboratory, a wish fulfilled paper on the role of the difficult word feeling in scien-
in 1912, but also had helped to found the British Jour- tific psychology. In 1931 he was elected to a chair in
nal of Psychology in 1904. Bartlett wrote an account experimental psychology at Cambridge. During his
(1937) of the early history of the Cambridge laborato- term as laboratory director, the number of faculty
ry. members grew and an increasing number of students
graduated in experimental psychology. An indication
When World War I began in 1914, Myers ap- of the success of the program was that, of the sixteen
pointed Bartlett relief director of the laboratory. professorships of psychology in Great Britain in 1957,
Bartlett instigated research into a variety of topics, in- ten were held by students of Myers and Bartlett. From
cluding studies of the detection of faint sounds, a 1922 to the beginning of World War II, Bartlett con-
project in which he collaborated with Emily Mary tinued to study conventionalization and memorizing
Smith, whom he married in 1920, and studies of indi- and wrote three books that show his interest in ap-
vidual differences in how subjects described pictures. plied psychology: Psychology and the Soldier (1927),
These individual differences, Bartlett believed, re- which dealt with personnel selection and war neuro-
flected above all subjective interests and socially de- ses, among other topics; Psychology and Primitive Cul-
termined interpretations: He ascribed the latter to ture (1923), in which he stressed the similarities rather
conventionalization; over the next few years he fo- than the differences between people in different so-
cused on the role of conventionalization in percep- cieties; and The Problem of Noise (1934).


always be at the mercy of old habits; but with a schema

one could revive individual memories that had been
laid down at widely varying periods of time and from
them form new combinations. He believed that con-
sciousness had evolved for this purpose, the looking
at or turning round on ones own schemata; the
ability to do this was greatly aided by the use of visual
images in addition to speech memory.
This insight could not be gained, argued Bartlett,
from studies of rote memory along the lines of Her-
mann Ebbinghauss experiments; further, the schema
theory allowed closer ties to be formed between ex-
perimental psychology and social psychology. Sche-
mata, he believed, were linked by appetites, in-
stincts, interests, and ideals, the first two laid down
particularly in childhood, the latter two in later life.
At the end of the twentieth century memory research-
ers still used the word schema, though some criticisms
had been raised about Bartletts terms schema and re-
construction (reviewed in Zangwill, 1972; Zangwill pre-
ferred the term abstraction).
Bartlett was made a fellow of the Royal Society in
1932, the year Remembering was published. His later
career was mainly devoted to applied psychology. He
was director of the Applied Psychology Unit at Cam-
bridge from 1944 to 1953; he was appointed to the
order of Commanders of the British Empire in 1941;
and he was knighted in 1948. His best-known work
during this period concerned fatigue states following
extended periods of skilled work. His book The Mind
The work for which Bartlett is best known is Re- at Work and Play (1951) is unusual in that it was in-
membering (1932), an elaboration of his research on tended for a juvenile audience. In Thinking (1958) he
conventionalization. In it he describes how he used discussed the development of the schema theory as an
two experimental paradigms to study memory: the example of scientific thinking. He died on September
method of repeated reproduction, in which a partici- 30, 1969.
pant studied a story or a picture and then reproduced Bibliography
it several times over a period of weeks or months; and Bartlett, F. C. (1923). Psychology and primitive culture. London: Cam-
the method of serial reproduction, in which a partici- bridge University Press.
pant recalled a story or a picture, then passed this (1925). Feeling, imaging, and thinking. British Journal of
production on to a second participant, who studied it, Psychology 16, 1628.
(1927). Psychology and the soldier. London: Cambridge Uni-
and so on down a chain of participants. Bartlett ob-
versity Press.
served that the two methods yielded similar results: (1932; reprint 1964). Remembering: A study in experimental
Recall was not duplicative but represented a recon- and social psychology. London: Cambridge University Press.
struction of the original story or picture based on (1934). The problem of noise. London: Cambridge University
memories of key details; the reconstruction could be Press.
(1936). Autobiography. In C. Murchison, ed., History of psy-
biased by conventionalization and importation.
chology in autobiography, Vol. 3. Worcester, MA: Clark Universi-
Given the fact that recall was not simply a duplica- ty Press.
(1937). Cambridge, England: 18871937. American Journal
tion of the same pattern over and over again, Bartlett, of Psychology 50, 97110.
following the suggestion of his neurologist friend (1951). The mind at work and play. London: Allen and Unwin.
Henry Head, argued that memories were not stored (1958). Thinking: An experimental and social study. London:
as static traces waiting to be revived; instead they Allen and Unwin.
formed parts of large complexes, called schemata, in Broadbent, D. E. (1970). Obituary of Sir F. C. Bartlett. Biographical
Memoirs of Fellows of the Royal Society 16, 116.
which individual components could be changed any Crampton, C. (1978). The Cambridge school: The life, works, and influ-
time there was a retrieval act. He argued that if traces ence of James Ward, W. H. R. Rivers, C. S. Myers, and Sir Frederic
were lifeless entities waiting to be revived, we should Bartlett. Ph.D. diss., University of Edinburgh, Scotland.

Harris, A. D., and Zangwill, O. L. (1973). The writings of Sir Fred- functions through techniques of introspection. Wat-
eric Bartlett, C.B.E., F.R.S.: An annotated handlist. British son boldly rejected this, asserting that behavior, per
Journal of Psychology 64, 493510.
Saito, A., ed. (1999). Bartlett: Culture and cognition. New York: Rout-
se, is the proper domain of psychology. For Watson,
ledge. prediction and control of overt behavior, rather than
Zangwill, O. L. (1972). Remembering revisited. Quarterly Journal of introspection of mental processes, formed the basis
Experimental Psychology 24, 123138. for an objective, scientific psychology. Behavior was
David J. Murray to be analyzed into stimulus-response (S-R) units
without appeal to hypothetical activities of brain or
mind. The units could be of widely varying size, from
the relatively molecular eyeblink elicited by a flash of
light to the more molar shopping trip as response
to an empty cupboard. Watson emphasized the conti-
See: AMNESIA, ORGANIC; GUIDE TO THE nuity between human and nonhuman species, and he
ANATOMY OF THE BRAIN stressed the importance of learning, in animals as well
as in humans, as the fundamental basis for under-
BASAL GANGLIA standing psychological process.

See: GUIDE TO THE ANATOMY OF THE BRAIN A neobehaviorism that came to the fore in the
1930s, that of Clark L. Hull and his student Kenneth
Spence, dominated until mid-century. Like Watson,
Hull described behavior as composed of S-R units,
but whereas Watson had presented S-R analyses as ad-
BEHAVIORISM justable in scale, the Hull-Spence approach focused
Most generally, behaviorism is a viewpoint that takes on molecular building blocks that were described as
psychological phenomena as physical activity rather forming chains of connecting events between envi-
than as belonging to a special domain of mental ronmental stimuli and observed behavior. These me-
events. For a behaviorist, then, psychology is the diating events included hypothetical (but presumably
study of behavior and its physical, mainly environ- physical) stimulus traces, covert responses, and re-
mental, determinants rather than of the nature of ex- sponse-produced stimuli. Learned S-R units were
perience or of mental process. Behaviorism originat- called habits. Hull contrived an elaborate theory
ed in natural-science traditions of the late nineteenth whose theorems and postulates, presented in geome-
century, and precursors of its methods and concepts ters style, were concerned with the formation of habit
developed at the turn of the century in the work of E. strength and with the mechanistic conversion of habit
L. Thorndike and Russian physiologist I. P. Pavlov, as strength into overt action. The theory was published
well as of several other psychologists and physiolo- as essentially complete in 1943. Although highly tout-
gists (Day, 1980; Herrnstein, 1969). ed, it proved ponderous, with numerous terms that
But behaviorism as a distinct viewpoint came to were difficult to evaluate; it fell of its own weight with-
be recognized with the publication of American psy- in a decade. Nevertheless, Hullian students gained
chologist John B. Watsons article Psychology as the dominant positions within academic psychology, and
Behaviorist Views It (1913). Identification of behav- elements of that approach can be discerned to this
iorism with the controversial Watson persists despite day in theorizing that rests on the metaphor of me-
the fact that it developed into several distinct tradi- chanical associative connections. Hulls emphasis on
tions that bear only a family resemblance to Watsons formal hypothesis testing, directed at hypothetical
views and to each other (Malone, 1990; Zuriff, 1985). constructs that are anchored to observable events as
The leading contemporary behaviorist position de- specified by operational definitions, also survives as
rives from the work of B. F. Skinner, which differs a methodological behaviorism (Skinner, 1945) that
from other behaviorisms in its detailed account of ver- has permeated much of psychology.
bal functioning and in its inclusion of activities such A counterpoint to Hulls views in the 1930s and
as thinking and feeling as behavior to be accounted 1940s was provided by Edward C. Tolman, who at-
for, while maintaining a primary focus on behavior- tempted to include purposive, intentional language
environment relations rather than upon processes in- within a behavioristic system. He invoked terms like
ferred as underlying those relations. purpose, expectation, and cognition to capture the larger-
Behaviorism originated in opposition to an or- scale, goal-oriented molar organization of behav-
thodox psychology that attempted to analyze con- ior. Tolman asserted that these terms need not imply
scious experience by focusing upon reports by observ- anything nonphysical or mentalistic; indeed, he em-
ers who were trained to examine their own mental ployed them in accounting for behavior of laboratory

rats as well as of humans. But Tolman undermined tion, examining its properties and its broad implica-
such disclaimers by characterizing his view as S-O-R tions. Empirically, he asked what would happen if
theory, with the O denoting a special role for pro- only some occurrences of a response are reinforced;
cesses within the behaving organism. The learning of he devised schedules of reinforcement to explore the
complex relationships, often characterized as cogni- many ways in which this can happen and their effects
tive maps, was said to mediate between environment on rates, patterning, and persistence of behavior.
and behavior. Critics of Tolmans account suggested Contemporary research has extended this to examine
that it left the organism buried in thought. To the issues such as the conditions of self-control and pref-
extent that he addressed the sources of action, Tol- erences among schedules that are relevant to microe-
man placed them within the organism, which tended conomics (e.g., Rachlin, 1989) and to biological theo-
to link his account with traditional mentalistic expla- ries of foraging (e.g., Fantino and Abarca, 1985).
nations of action. Thus it is not surprising that Tol- Interpretatively, Skinner addressed the functional
mans inclusion of intentional language never was ac- characteristics of verbal behavior, describing how an
cepted by the broader behavioristic community. individual affects the behavior of others and how oth-
B. F. Skinner also departed from the S-R behav- ers teach the individuals verbal discriminations
ioral mainstream of the 1930s and 1940s, but in dif- (Skinner, 1957). His approach initially was not wel-
ferent ways. He rejected mentalistic terms as mis- comed by linguists, but later developments in linguis-
leading fictions while including the relationships tics are more congenial to it (Andresen, 1990). The
that were Tolmans primary concern. Skinners first analysis includes activities like thinking, feeling, and
conceptual innovation was to reformulate the reflex; even introspecting as behavior to be accounted for
he described this simplest unit of behavior not as rather than as special bases for explaining overt ac-
stimulus-response connection but as directly observ- tion. It asserts that individuals know their private
able abstraction, a correlation between classes of stim- thoughts and feelings less well than they know exter-
uli and classes of responses. Then he distanced his nal events, because the world cannot as accurately
theory further from mediational notions of mecha- teach individuals to discriminate the former (Skinner,
nism and associative connection by delineating non- 1963). This provocative position gains independent
reflexive behavioral units that he called operants. Op- support from the philosophies of Ryle (Schnaitter,
erants act upon the environment; they are selected by 1985) and Wittgenstein (Day, 1969).
their consequences through processes denoted as re- Skinner also addressed ethical and social issues in
inforcement, punishment, and extinction. Operants can light of reinforcement-based principles, speculatively
range from small to large, and are defined by the con- in Walden Two, a utopian novel that sketches an ex-
sequences that shape or maintain them but also by the perimental approach to communal living, and analy-
contexts within which the selecting consequences tically in essays such as The Ethics of Helping Peo-
have occurred. The result is a three-term relationship ple (Skinner, 1978), which asserts that human rights
composed of classes of responses, consequences, and properly concern empowerment of effective action
discriminative stimuli. rather than access to things or services. Extensive dis-
Operant behavior often is characterized in ordi- cussions of these and other implications of reinforce-
nary language as intentional and purposive, thus hav- ment theory are provided by Skinner (1971, 1974),
ing the molar characteristics that were Tolmans and by Catania and Harnad (1988). Contemporary
primary concern. But the traditional appeal to men- behavioral research related to language emphasizes
talistic intention is replaced by environment-based se- relationships between verbal and nonverbal behavior
lection in this account of action, just as in Darwinian (Cerutti, 1989; Hayes, 1989) and issues such as the
biology natural selection replaces divine intention in nature and origins of symbolic functioning (Sidman,
the account of new species. Learning of new behavior 1986).
is readily demonstrated by rapidly shaping new pat- Of contemporary approaches, the most distinctly
terns through differential reinforcement and through behavioral one is behavior analysis. Extending from
gradual fading of discriminative stimuli. In later
Skinners work, it differs philosophically and concep-
work, Skinner (1984) described selectionist principles
tually from other behaviorisms as well as from main-
as applying to behavior patterns at the evolutionary
stream psychology (Lee, 1988). Its pragmatic contri-
level as well as at the level of cultural practice, giving
butions have proved effective in such diverse settings
accounts with close affinity to contemporary work in
as health-maintenance programs concerned with
anthropology (e.g. Lloyd, 1985) and biology
weight control, smoking, and wearing of automobile
(Dawkins, 1982; Smith, 1986).
seat belts; frequent flier marketing techniques
While his theory also included other principles, launched by airlines; techniques for basic research on
Skinner emphasized reinforcement as a basic rela- drug addiction as well as for its treatment; education-

al techniques of documented effectiveness for handi- (1945). The operational analysis of psychological terms.
capped and disadvantaged, as well as for mainstream, Psychological Review 52, 270277, 291294.
(1948). Walden two. New York: Macmillan.
children (Becker, 1978; Wolf et al., 1987); and inno- (1957). Verbal behavior. New York: Appleton-Century-
vative formats for personalized instruction at the col- Crofts.
lege level (Keller, 1977). Contemporary behaviorists (1963). Behaviorism at fifty. Science 134, 566602.
are represented by professional organizations that in- (1971). Beyond freedom and dignity. New York: Knopf.
clude several thousand researchers, scholars, and (1974). About behaviorism. New York: Knopf.
(1978). Reflections on behaviorism and society. Englewood
practitioners (Thompson, 1988) whose work is repre- Cliffs, NJ: Prentice-Hall.
sented by more than a score of primarily behavioral (1984). Selection by consequences. Science 213, 501504.
journals (Wyatt et al., 1986). Thus, behaviorism has Smith, T. L. (1986). Biology as allegory: A review of Elliott Sobers
extended well beyond, while continuing an apposi- The nature of selection. Journal of the Experimental Analysis of Be-
tional role within, the specialized field where it began. havior 46, 105112.
Thompson, T. (1988). Benedictus behavior analysis: B. F. Skinners
See also: OPERANT BEHAVIOR; SKINNER, B. F.; magnum opus at fifty. Contemporary Psychology 33, 397402.
Tolman, E. C. (1932). Purposive behavior in animals and men. New
York: Appleton-Century-Crofts.
Watson, J. B. (1913). Psychology as the behaviorist views it. Psycho-
logical Review 20, 158177.
Andresen, J. T. (1990). Skinner and Chomsky thirty years later. Hi- (1919). Psychology from the standpoint of a behaviorist. Philadel-
storiographia Linguistica 17, 145165. phia: J. B. Lippincott.
Becker, W. C. (1978). The national evaluation of follow-through: Wolf, M. M., Braukmann, C. J., and Ramp, K. A. (1987). Serious
Behavior theory-based programs come out on top. Education delinquent behavior as part of a significantly handicapping
and Urban Society 10, 431458. condition: Cures and supporting environments. Journal of Ap-
Catania, A. C., and Harnad, S. (1988). The selection of behavior, the plied Behavior Analysis 20, 347359.
operant behaviorism of B. F. Skinner: Comments and consequences. Wyatt, W. J., Hawkins, R. P., and Davis, P. (1986). Behaviorism: Are
New York: Cambridge University Press. reports of its death exaggerated? The Behavior Analyst 9, 101
Cerutti, D. T. (1989). Discrimination theory of rule-governed be- 105.
havior. Journal of the Experimental Analysis of Behavior 51, 257 Zuriff, G. E. (1985). Behaviorism: A conceptual reconstruction. New
276. York: Columbia University Press.
Dawkins, R. (1982). The extended phenotype. San Francisco: Freeman.
Day, W. F. (1969). On certain similarities between the Philosophical Philip N. Hineline
investigations of Ludwig Wittgenstein and the operationism of
B. F. Skinner. Journal of the Experimental Analysis of Behavior 12,
(1980). The historical antecedents of contemporary behav-
iorism. In R. W. Rieber and K. Salzinger, eds., Psychology: The- BEHAVIOR THERAPY
oretical-historical perspectives, pp. 203262. New York: Academ-
ic Press. Behavior therapy is a term used to describe a number
Fantino, E., and Abarca, N. (1985). Choice, optimal foraging, and of therapeutic procedures that share certain assump-
the delayreduction hypothesis. Behavioral and Brain Sciences tions about the nature of behavioral and psychologi-
8, 315-362.
Hayes, S. C., ed. (1989). Rule-governed behavior: Cognition, contingen- cal problems and how they can best be overcome. The
cies, and instructional control. New York: Plenum. procedures can be classified into three main groups:
Herrnstein, R. J. (1969). Behaviorism. In D. L. Krantz, ed., Schools fear-reduction procedures, operant conditioning
of psychology: A symposium, pp. 5168. New York: Appleton- procedures, and aversive techniques.
Hull, C. L. (1943). Principles of behavior. New York: Appleton- The major fear-reduction procedures consist of
Century-Crofts. systematic desensitization, in which the person is trained
Keller, F. S. (1977). Summers and sabbaticals: Selected papers on psy- to imagine a series of increasingly fearful images
chology and education. Champaign, IL: Research Press. while in a state of relaxation; therapeutic modeling, in
Lee, V. (1988). Beyond behaviorism. Hillsdale, NJ: Erlbaum.
Lloyd, K. E. (1985). Behavioral anthropology: A review of Marvin which the person observes and then imitates a thera-
Harriss Cultural materialism. Journal of the Experimental Analysis pist model engaging in increasingly close contact with
of Behavior 43, 279287. the frightening object or situation; and flooding, in
Malone, J. C. (1990). Theories of learning: A historical approach. Bel- which the phobic person is exposed to intensely fear-
mont, CA: Wadsworth. ful stimulus situations for prolonged periods. In all of
Rachlin, H. (1989). Judgment, decision, and choice: A cognitive/
behavioral synthesis. New York: W. H. Freeman. these methods, the phobic person is exposed to the
Schnaitter, R. (1985). The haunted clockwork: Reflections on Gil- real or imagined fear stimulus repeatedly and/or for
bert Ryles The concept of mind. Journal of the Experimental Analy- prolonged periods, and in all of them attempts to es-
sis of Behavior 43, 145153. cape from or avoid the fear stimulus are discouraged
Sidman, M. (1986). Functional analysis of emergent verbal classes. (response prevention). The combination of exposure
In T. Thompson and M. D. Zeiler, eds., Analysis and integration
of behavioral units, pp. 213245. Hillsdale, NJ: Erlbaum.
and response prevention has proved to be a robust
Skinner, B. F. (1938). The behavior of organisms: An experimental anal- and dependable means for reducing fear, and the
ysis. New York: Appleton-Century-Crofts. clinical efficacy of this combination in each of the

three forms of fear-reduction procedures has been desensitization remains useful for numerous prob-
confirmed in numerous controlled clinical trials lems, especially those in which direct exposure is im-
(Marks, 1987; OLeary and Wilson, 1987). practical or unacceptable, as in the treatment of cer-
tain sexual disorders and social phobias.
All three methods can be traced back to the work
of the Russian physiologist Ivan Petrovich Pavlov on All of the fear-reduction techniques are applica-
conditioning, and especially to his research on exper- tions of learning procedures to clinical problems,
imental neurosis. In developing the first of the mod- and, in common with the other forms of behavior
ern methods, systematic desensitization, Wolpe therapy, are based on the assumption that most psy-
(1958) was influenced by the writings of Pavlov and chological problems can be overcome by the use of
the modern learning theorists, especially Clark Hull. conditioning or other learning processes. In the early
Having rejected the psychodynamic approach, Wolpe stages of behavior therapy, it was assumed that most
attempted to apply modern learning techniques to psychological problems are the result of faulty learn-
psychological problems, particularly those in which ing (for instance, Symptoms are unadaptive re-
anxiety is prominent. After completing a series of ani- sponses and Symptoms are evidence of faulty learn-
mal experiments, he concluded that graded, gradual ing [Eysenck and Rachman, 1965, p. 12]).
re-exposures to a fearful stimulus are the best way to Furthermore, it was argued that problems that are the
weaken or eliminate the fear. He also concluded that result of faulty learning can be unlearned. In due
the fear-reduction process can be facilitated by the course, more complex explanations were substituted.
deliberate superimposition on the evoked fear re- The second form of behavior therapy, consisting
sponse of a competing incompatible response (such of the application of operant conditioning ideas and
as relaxation imposed on a fear response). Each occa- procedures to clinical problems, was engineered in
sion on which an incompatible response is imposed the United States and applied mainly to the psycho-
over the fear response is an instance of reciprocal in- logical problems of children and adults with severe
hibition. Wolpe argued that repeated instances of handicaps (e.g., mentally retarded people in or out of
such reciprocal inhibition give rise to a relatively per- institutions and people with chronic and serious psy-
manent form of conditioned inhibition (of fear), and chiatric disturbances, especially chronic schizophre-
that the therapeutic effects are a direct consequence nia). The fear-reduction techniques, developed main-
of the reciprocal inhibition. ly in Britain, are still used predominantly in dealing
Desensitization is the earliest and best- with adult neurotic problems, especially anxiety dis-
established of the methods, but Wolpe introduced a orders such as obsessional disorders, panic disorders,
number of other therapeutic procedures. Most atten- social phobias, and circumscribed phobias.
tion, however, has been devoted to desensitization, The clinical application of operant conditioning
which has been the subject of considerable experi- later referred to as reinforcement therapy was a direct
mental work and testing. This research advanced the application of Skinnerian ideas, with emphasis on the
therapeutic efficacy of these results and gave rise to consequences of behavior. Behavior that is followed
a fresh view of fear itself. by a reward will be strengthened and behavior that is
Additions to and improvements of the clinical followed by nonreward will be weakened.
techniques were introduced in the early 1970s, and The first application consisted of a series of at-
for certain types of anxiety disorders (such as panic tempts to treat schizophrenic problems in laboratory
disorder, agoraphobia, obsessional disorders, simple settings, but only limited success was achieved until
phobias), therapeutic modeling replaced desensitiza- the methods were applied, often with considerable in-
tion as the method of choice. In most cases, therapeu- genuity, to the maladaptive behavior of institutional-
tic modeling and flooding are carried out in vivo, an ized patients with chronic psychiatric disorders (e.g.,
unfortunately chosen term that here means exposure Ayllon and Azrin, 1968). Notable advances were
to the fear stimulus rather than to an imagined repre- made by the selective reinforcement of desirable be-
sentation of the stimulus (as in the desensitization havior (e.g., self-care, eating) and the withholding of
procedure). The development of therapeutic model- (social) reinforcement after undesirable behavior.
ing, largely the result of Banduras work (1969), con- Many of these useful advances were later incorporat-
sists of repeated exposures to the fear stimulus. The ed into an institutional program that Ayllon and
phobic person first watches and then imitates the ap- Azrin, two pioneers of this work, called the token econo-
proach behavior of a therapeutic model. The method my (exchangeable tokens having replaced tangible re-
is effective and well accepted by most subjects, clients, wards). The fact that the large ambitions of the earlier
and patients. Flooding is seldom the first choice of workers who expected reinforcement therapy to elim-
treatment but can be used in certain cases such as ex- inate these psychiatric problems were never achieved
tensive obsessional/compulsive problems. Systematic does not detract from the contribution that this form

of therapy continues to make. It is widely used to OLeary, K., and Wilson, G. T. (1987). Behavior therapy, 2nd edi-
modify the maladaptive behavior of retarded chil- tion. Englewood Cliffs, NJ: Prentice-Hall.
Rachman, S. (1990). Fear and courage, 2nd edition. New York: W.
dren and adults, patients with chronic psychiatric dis- H. Freeman.
orders, children with speech or other behavioral defi- Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA:
cits, and in educational settings. Stanford University Press.
Aversion therapy is a direct application of Pavlovian Stanley J. Rachman
conditioning to appetitive but maladaptive behavior
such as the excessive use of alcohol and was originally
prompted by the discovery that laboratory animals
can develop conditioned nausea reactions to the stim- BIRDSONG LEARNING
uli that are associated with the administration of Song behavior refers to complex vocalizations used in
drugs that induce nausea. In order to convert the ab- the context of mate attraction and territorial defense.
errant stimuli (such as inappropriate sexual stimuli or Birds that produce such sounds are commonly called
alcohol) into conditioned stimuli for nausea or other songbirds. Technically songbirds constitute species in
unpleasant responses, the alcohol or sexual stimuli the avian order Passeriformes. This is by far the larg-
are contingently followed by aversive stimulation. est avian order and contains about half of the more
The earlier and still most common form of aversion than 9,000 living bird species. The songbird order,
therapy consists of pairing an alcohol-conditioned one of the most recently evolved, includes familiar
stimulus with aversive nausea induced by drugs. This avian groups such as sparrows, swallows, starlings, ca-
form of chemical aversion therapy is used mainly in naries, finches, warblers, jays, titmice, crows, wrens,
the treatment of alcoholism; the other technique, elec- robins, and buntings. This order can be further divid-
trical aversion therapy, in which the aversive stimulus is ed into two suborders, the Oscines (members of the
an electric shock, is used mainly in the treatment of suborder Passeres) and the sub-Oscines (a much smal-
aberrant sexual behavior such as pedophilia. Al- ler group that includes the flycatchers of North Amer-
though aversion therapy appears to make a useful ica), which appeared earlier in evolutionary history
contribution to dealing with alcohol and sexual prob- and is thought to be more primitive. All songbirds
lems, it is rarely sufficient and nowadays is used as produce complex vocalizations, but there do appear
part of a wider therapeutic program that typically in- to be qualitative differences between Oscine species
cludes counseling, group therapy, and family thera- and subOscine species in vocal development and in
py. This insufficiency, the ethical problems involved the neural substrate mediating vocal learning and
in the deliberate application of aversive stimuli, and production.
the fact that the precise nature of the conditioning
processes involved in aversion therapy is not fully un-
derstood have combined to limit its use. The Basics of Birdsong
The other two forms of behavior therapy also All songbirds have a repertoire of up to twenty or
have their limitations, so clinicians have been recep- so distinct vocal sounds that they use for communica-
tive to the growing influence of cognitive analyses of tion about danger, food, sex, group movements, and
psychological and clinical phenomena. Behavior for many other purposes. One can usually make a dis-
therapy has been expanded to include the influence tinction between a birds calls, which are usually brief
of cognitive factors, and most practitioners now favor and monosyllabic, and its songs, which are more ex-
cognitive behavior therapy, which combines the original tended patterns of sound and often tonal and melod-
forms of behavior therapy with cognitive analyses of ic. The decision to classify a vocalization as a song or
the problem and cognitive procedures for helping to a call is usually based on the perceived function. The
deal with them. Cognitive behavior therapy appears main functions that have been ascribed to song be-
to have achieved most success in the treatment of de- havior are territory defense (or spacing behavior) and
pression (Beck, 1976; OLeary and Wilson, 1987). mate attraction, as opposed to calls, which are in-
volved in such functions as signaling danger or food
Bibliography and maintaining flock cohesion. Songs, especially
Ayllon, T., and Azrin, N. (1968). The token economy. New York: Ap- among songbird species in the temperate zone, are
Bandura, A. (1969). Principles of behavior modification. New York:
usually produced by males. Among tropical species
Holt. females as well as males often sing. In some species,
Beck, A. (1976). Cognitive therapy and the emotional disorders. New males and females sing a coordinated song that is
York: International University Press. known as a duet.
Eysenck, H., and Rachman, S. (1965). The causes and cures of narco-
sis. London: Routledge and Kegan Paul. Unlike most calls, songs are learned: They devel-
Marks, I. (1987). Fears, phobias, and rituals. Oxford: Oxford Univer- op abnormally if a young male is reared out of hear-
sity Press. ing of the sounds of adults (see Figure 1). Among spe-

cies in the songbird order only the Oscines clearly to rely on learning for vocal development. The avian
learn their songs; in the few sub-Oscine species inves- groups known to have learned songs include hum-
tigated song learning does not appear to occur thus mingbirds, parrots, and all Oscine songbirds.
providing a natural comparison between closely relat-
ed species that vary in one key aspect of their vocal
behavior. A common consequence of this dependence Sensitive Periods and the Timing of Song
on learning among Oscines is the emergence of local Learning
song dialects, varying on much the same geographic There is an underlying pattern in the steps typi-
scale as dialects in human speech. These dialect cally required in learning to sing. First is the acquisi-
boundaries have been shown to serve as imperfect tion phase, when a bird hears songs and commits
barriers for gene flow in species such as the white- some to memory. These are stored for a period that
crowned sparrow indicating that this intraspecific varies in duration from species to speciesfrom days
geographic variation is meaningful (MacDougall- to monthsuntil the bird begins to recall songs from
Shackleton and MacDougall-Shackleton, 2001). memory and starts to produce imitations, sometimes
faithful to the original model, sometimes far differ-
Songbirds are unique among animals in the many
ent. Thus there is a separation in time between the ac-
analogies that can be struck between song learning
quisition or sensory phase and the production or sen-
and the acquisition of human speech (Doupe and
sorimotor phase, ending with the production of
Kuhl, 1999). Avian vocal development provides one
crystallized, adult song (Marler, 1997).
of the few tractable animal models for studying the
behavioral, hormonal, and neural bases of vocal plas- There are often sensitive periods for song acquisi-
ticity (Ball and Hulse, 1998). No known nonhuman tion, sometimes restricted to a short period early in
primate depends upon learning to develop its natural life, and sometimes extending into adulthood. Even
vocal repertoire. Other than humans, cetaceans and close relatives, such as sparrows and canaries, may dif-
perhaps some bats are the only mammals that appear fer in this respect. Several species of sparrows have a

sensitive period for acquiring songs beginning at or guided learning. In several sparrow species, for
about twenty days of age, soon after young males be- example, species-typical vocalizations appear to be
come independent from their parents, and ending privileged because they are learned preferentially.
four to six weeks later (Nelson, 1997). Such sensitive Such preferences have even been observed for specif-
periods for learning are variable within limits, de- ic subspecies (Nelson, 2000). Another aspect of this
pending on the strength of song stimulation and the memory, strongly advocated by Peter Marler and col-
influence of physiological factors, such as hormonal leagues, is that the memory encodes species-typical
states, that vary with the season. If young are hatched patterns of vocal behavior even before it hears the
late in the season and singing has already ceased for song of its own species. This idea of an innate specifi-
that year, closure of the sensitive period may be de- cation of species-typical vocalizations would be one
layed until the following spring. The experimental way to explain why even in isolated-reared sparrows
withholding of stimulation by songs of the birds own many species-typical attributes of their vocalizations
species can also delay closure of the learning (Kroods- are still apparent in the abnormal songs produced by
ma and Pickert, 1980; Baptista and Gaunt, 1997). these birds. According to this view of song learning,
Species that learn song only during a sensitive period the formation of the auditory memory guiding song
early in ontogeny are known as age-limited learners. learning would result from a memorization by selec-
Species that continue to learn new songs throughout tive activation and attrition of innate circuits rather
their lives (such as canaries and European starlings) than from a selective instruction process (Marler,
are referred to as age-independent learners. In addi- 1997). Researchers have not yet resolved the relative
tion, there are species that fall in between these ex- validity of these two opposing models of auditory
treme cases in terms of song development. memory formation.

Effects of Isolation and Deafness Song Overproduction, Social Interactions,

Regardless of whether or not they have had the and Action-Based Learning
chance to learn, songbirds can always produce some
The young male songbird typically begins singing
aspects of the normal song of their species. When
sometime after he has memorized learned songs, but
sparrows are raised in isolation, for example, the note
the imitations are not fully formed. Instead, the
structure and tonal quality of their songs is abnormal,
young male starts with subsong, an amorphous, noisy
but each species still produces some basic features of
twittering that changes gradually, first into plastic
normal song syntax (see Figure 1). These features
song, which also is highly variable but contains the
are produced irrespective of whether they were repre-
first obvious signs of mature song structure, finally
sented in any songs a male may have learned.
crystallizing into the stable patterns of mature, adult
Insight into the basis of this ability to produce song (see Figure 2). Subsong resembles the babbling
certain normal song features is gained by studying the of human infants and may be important for develop-
songs of deaf birds, which are highly abnormal and ing the motor skills of singing and other prerequisites
variable in structure (see Figure 1). This degraded for song learning, such as the ability to guide the
form of singing is observed both if a male becomes voice by the ear. Rehearsal of previously memorized
deaf before song stimulation and if he is deafened song patterns begins in plastic song. Often more plas-
after song stimulation but before the development of tic song themes are produced than are needed for
singing. There seems to be no internal brain circuitry mature singing (see Figure 3), and many are discard-
that makes memorized songs directly available to ed when song crystallization occurs. Interactions be-
guide motor development. To transform a memo- tween a male and his neighbors can influence which
rized song into a produced song, the bird must be of the overproduced songs will be selected for crystal-
able to hear its own voice. One can infer that there are lization and inclusion into the final repertoire (Nel-
auditory memories for song, involved in guiding song son, 1997). Songs most similar to the neighbors
production, conceived of as neural mechanisms in the songs are retained; those that are different are reject-
auditory circuitry of the songbird brain that must vary ed. It is as if a process akin to operant conditioning
in their specifications from species to species (Kon- influences which songs are selected for later crystalli-
ishi, 1965; Marler, 1997). zation among those in auditory memory. Marler and
Douglas A. Nelson (1993) refer to this process as ac-
tion-based learning to contrast it with memory-
The Nature of the Auditory Memory based learning. The latter refers to cases where crys-
Guiding Song Learning tallized songs are produced in reference to previously
One aspect of the process of memory formation formed memories independently of social interac-
needed for song learning is that it involves selective tions during the sensorimotor period.

then to nXIIts constitutes a motor pathway special-

ized for song production. RA also projects to brain-
stem areas controlling respiration, which needs to be
coordinated with song (Wild, 1997). The anterior
forebrain song nuclei lMAN (the lateral part of the
magnocellular nucleus of the anterior neostriatum)
and area X, as well as the thalamic nucleus that inter-
connects them, DLM (the medial portion of the dor-
solateral nucleus of the thalamus), form a second
pathway connecting HVc to RA. All of the song nuclei
are present bilaterally, and the thalamic song nucleus
Uva (the uvaeform nucleus) forms a connection be-
tween the two sides.
The presence of a song system is particularly as-
sociated with the learning of song, not just with song
production. These brain areas are found only in birds
that sing and that learn their song by reference to au-
ditory information. For example, sub-Oscine song-
birds that do not learn their vocalizations do not seem
to have telencephalic song nuclei such as HVc or RA.
The vast majority of vocal learners are Passerine
songbirds, but evolutionarily unrelated species that
are vocal mimics, such as parrots or hummingbirds,
which require conspecific learning to produce their
complex vocalizations, are also found to have fore-
brain structures resembling nuclei in the song system
including HVc and RA (Jarvis and Mello, 2000; Jarvis
et al., 2000; Gahr, 2000).

A Motor Pathway for Song

Evidence for a specialized motor pathway for
song comes from three sources. Behavioral studies
show that lesions of HVc, RA, or the hypoglossal
nerve cause severe disruption of adult song. Also,
electrophysiological recordings in Nif, HVc, and RA
of awake birds reveal neurons whose activity is highly
correlated with singing (Yu and Margoliash, 1996).
These data indicate a hierarchy in motor processing
with activity in HVc correlated with syllable produc-
tion and activity in RA correlated with note produc-
The Song System tion (Yu and Margoliash, 1996). Nif may be the bird-
A discrete set of brain areas discovered by F. Not- song pattern generator, because section of the tract
tebohm (Nottebohm, 1996), often called the song sys- from Nif to HVc destroys the patterning of song,
tem, controls song learning and production. The while lesions of Uva do not (McCasland, 1987).
song system suggests a possible location for the neural Motor-driven immediate early gene induction also
changes associated with song learning. The song sys- occurs in nuclei such as HVc and RA in that song pro-
tem consists of numerous interconnected nuclei and duction is correlated with such induction in deaf birds
occupies a relatively large volume of brain (see Figure induced to sing (Jarvis and Nottebohm, 1997).
4). The brain area nXIIts (the tracheosyringeal por- The song motor commands must ultimately be
tion of the hypoglossal nucleus) contains the motor translated into a pattern of syringeal muscle activa-
neurons that control the musculature of the birds tion and respiratory control. The output nucleus of
vocal organ, the syrinx. The pathway from Nif (the the song system, nXIIts, has a map of the muscles of
nucleus interface) through HVc (or high vocal center) the syrinx: All motor neurons projecting to a particu-
to RA (the robust nucleus of the archistriatum) and lar syringeal muscle are clustered together in discrete

Figure 3

Diagram illustrating the process of song development in the male swamp sparrow. The horizontal axis represents the birds age from
hatching in months; the successive rows in the vertical axis represent the different processes involved in establishing the adult
crystallized song and how these processes were experimentally investigated. A male successively exposed to different song types at
various ages (top row, A through G) will only memorize song types presented during the sensory acquisition phase. After a period of
subsong during which no song type is recognizable, the bird will start producing during the period of plastic song all song types that
were stored in memory. Selective attrition will then take place, in part as a consequence of social interactions with neighbors, and
crystallized adult song will then only include one or two selected song types.

zones within the motor nucleus. RA also has zones of two separate sources are used to produce a species-
premotor neurons that project to the corresponding typical vocalization. Thus initial claims that control of
muscle control area in nXIIts. The dorsal portion of song production is lateralized similarly to speech pro-
RA does not project to nXIIts but to the dorsomedial duction is an oversimplification. It seems that asym-
nucleus (DM) of the midbrain, an area thought to be metries in syringeal control when they do occur re-
involved in vocalization and respiration in all birds, flect a peripheral asymmetry rather than a
including nonsongbirds (see Figure 4; Vicario, 1991). hemispheric specialization, as is the case in humans.
Neurons in DM then project to nXIIts, suggesting
that RA has two parallel and perhaps functionally dif-
Song-Selective Auditory Pathways and
ferent inputs to the syringeal motor neurons.
Feedback Mechanisms
Intrabronchial measurements of both airflow and Because auditory feedback is used to correct vocal
sound from each syringeal side have clarified how the motor output during song learning, there must be a
syrinx functions to produce sound (Suthers, Goller, link between the auditory system and the vocal motor
and Pytte, 1999). The syrinx contains two separate pathway. In addition, there must be brain mecha-
sound sources, each with an independent motor con- nisms for very specific recognition of song and song-
trol. There is extensive species variability in how these like vocalizations. Auditory information is relayed to

Figure 4

Schematic representation of the song system in Oscines. The diagram illustrates many of the known nuclei and connections in the
song system. Different fill patterns have been used to highlight nuclei that are part of the motor pathway (black), the anterior
forebrain pathway mostly involved in song learning (cross-hatching) and the parts of the auditory pathway that convey auditory
information to nuclei of the song system (white).

the songbird forebrain as in all other birds, traveling auditory stimuli. Some of these HVc auditory neurons
from the cochlear nuclei through the thalamus to the are song-selective: They respond best of all to the
forebrain primary auditory area, Field L. Anatomical birds own song, even in comparison with very similar
experiments show that subdivisions of Field L proj- songs of conspecifics (Margoliash, 1986). HVc song-
ects to the vicinity of HVc and RA (Vates, Broome, selective neurons are also sensitive to temporal order:
Mello, and Nottebohm, 1996; see Figure 4). Howev- They are activated more strongly by the birds own
er, electrophysiological investigations also suggest song when the syllables are in the normal sequence
that auditory information gains access to HVc via Nif than when the identical song components are played
(Janata and Margoliash, 1999). Field L projects to the out of order or in reverse. This high degree of selec-
caudolateral ventral hyperstriatum (clHV) that in tivity in individual neurons provides a possible mech-
turn projects to Nif (Vates, Broome, Mello, and Not- anism for specific recognition of song. Because the
tebohm, 1996). In addition to song-related motor birds own song is learned during development, this
neurons, HVc also contains neurons that respond to song selectivity must also be learned. In fact, studies

in young birds have shown that these neurons acquire learning did not require slow NMDA-EPSCs at syn-
their specificity during sensorimotor learning. apses critical for song development.
Lesion studies have shown that the anterior fore-
brain pathway containing MAN and X (see Figure 4) The Development of the Song System,
plays an important role in song learning but is not an Hormonal Effects, and Sex Differences in
essential part of the motor pathway for song in the Brain and Behavior
adult. Lesions of MAN or X have no apparent effect It is a striking feature of the song system that it
on adult song production, but destruction of either of continues to develop after hatching, during song
these areas in young birds results in markedly abnor- learning. Administration of 3H-thymidine can be
mal song (Bottjer and Johnson, 1997). There are used to label neurons undergoing their last cell divi-
song-selective auditory neurons, similar to those in sion, or birthdate. Such birthdating of song nuclei
HVc, in every nucleus in this loopX, DLM, and shows that MAN and RA are born before hatching,
MAN. Like the neurons in HVc, these auditory neu- but that there is significant neurogenesis in HVc and
rons acquire their song selectivity in parallel with X in the first several months after hatching (Alvarez-
song acquisition. If the syrinx is selectively dennervat- Buylla and Kirn, 1997). There is also naturally occur-
ed prior to the sensorimotor phase of song learning, ring cell death during postnatal development. In
male zebra finches in many cases are unable to match male zebra finches, many MAN neurons die around
their vocal output to the tutors song (Solis and five weeks after hatching. Synaptic connectivity con-
Doupe, 2000). In such birds, many neurons in area X tinues to develop at these late stages. The motor pro-
exhibited a dual selectivity and responded equally jection from HVc reaches its target nucleus RA by
well to the birds own song and to the tutor song. The postnatal day fifteen, but then waits outside RA for
degree of selectivity for these stimuli as compared to about ten days before growing in and completing the
conspecific song or reversed song was considerably circuit (Konishi and Akutagawa, 1985). Interestingly,
less than in normal adults (Solis and Doupe, 2000). male zebra finches first begin to sing at the time of in-
growth of HVc axons into RA. Connections from HVc
One possible function of this song-selective audi-
to MAN via X and DLM are present and functional
tory pathway is to act as a correction signal provided
by day fifteen, but topographic features of the projec-
by auditory feedback for the learning of song and the
tion from lMAN to RA occur at days twenty to twenty-
maintenance of learned song. There is evidence that
five during the early stages of vocal learning. Timing
the delayed auditory feedback can cause a gradual de-
of this projection is delayed if juvenile birds are de-
terioration of adult zebra finch song (Leonardo and
prived of normal conspecific auditory experience
Konishi, 1999). Deafening adult male zebra finches
(Iyengar and Bottjer, 2002).
also results in a deterioration of song (Nordeen and
Nordeen, 1992). However, lesioning a nucleus in the Bird song can vary among the sexes. Typically in
anterior forebrain pathway such as lMAN prevents temperate zone species, male birds sing for courtship
the deterioration in song resulting from deafening and territorial defense, while female birds sing much
(Brainard and Doupe, 2000). These findings suggest less or not at all. In the tropics the pattern is quite dif-
that an active process is required for the maintenance ferent: males and females often stay together on terri-
of song and that the anterior forebrain pathway is es- tories all year round, and both sexes will sing to de-
sential to this process. fend these territories. The behavioral dimorphism
between males and females has a striking correlate in
The song-selective auditory loop from HVc to the the sexual dimorphism of the song system itself. Ini-
anterior forebrain eventually projects back into the tial studies of species the exhibit extreme differences
motor pathway through its connection to RA (see Fig- in song behavior: For example, zebra finches and ca-
ure 4). This convergence of auditory and motor in- naries revealed marked male-biased differences be-
puts makes RA a possible site for the auditory guid- tween males and females (Nottebohm and Arnold,
ance of vocal motor development during learning. 1976). Nottebohm and Arnolds discovery provided
The NMDA subclass of glutamate receptors is an opportunity to explore sex differences in the brain
thought to play a role in some forms of synaptic plas- in a truly comparative sense. Comparative studies
ticity. NMDA receptor-mediated EPSCs (NMDA- have been employed to understand the function of
EPSCs) become fast during song development, a tran- these sex differences in the brain (see Figure 5). In
sition posited to limit learning. However, manipula- some songbird species, females sing rarely or not at
tions of the sensitive period for song learning by all, and the brain nuclei that control song are many
isolating nestling zebra finches delayed NMDA- times larger in volume in males than in females. In
EPSCs but did not prevent the birds from learning other species, males and females sing approximately
(Livingston, White, and Mooney, 2000). Thus song equally, and the brain nuclei that control song are ap-

Figure 5

Interspecific variations in the volumetric sex difference of the song control nucleus HVc in songbirds and their relationship with the
sex difference in singing behavior. The figure illustrates the ratio of the HVc volume in males and females in species where the female
normally never sings (very large HVC in males compared to females), where females sing but less than males (intermediate ratio of
HVc volumes), and where females sing about as often as males (duetting species, ratio of HVc volumes is close to 1). The volume ratio
in three species of the first group has been indicated as approximately infinite () because HVc cannot be detected in females.

proximately equal between the sexes. Recently statis- Steroid hormones influence both singing and the
tical methods have been employed control for phylo- song system. Singing is much increased when andro-
genetic effects while comparing the coevolution of gen levels are high, for instance during the breeding
traits. This analysis indicates that the evolution of sex season in the spring. In several songbird species,
differences in song has coevolved with the evolution there is also a seasonal variation in the size of the song
of sex differences in singing behavior in songbird nuclei (Tramontin and Brenowitz, 2000). In canaries,
species (Figure 5; MacDougall-Shackleton and Ball, where researchers first described this trait, HVc and
1999). It is unclear whether these morphological RA enlarge by approximately 50 percent in the spring
differences related to variation in song are just (Nottebohm, 1981). These seasonal changes in vol-
related to differences in production or if differences ume involve changes in cell size and cell number
in song learning also occur. Sex differences in (Tramontin and Brenowitz, 2000). These seasonal
volume in nuclei such as HVc and RA are asso- morphological changes were initially thought to be
ciated with differences in cell size and cell num- related to seasonal changes in song learning but cur-
ber. Other attributes of the phenotype of cells in rent data support the notion that the changes are
these nuclei are different in males and females, such more closely related to seasonal changes in song per-
as the number of cells expressing androgen recep- formance (Tramontin and Brenowitz, 2000). In some
tors. species, such as canaries and white-crowned sparrows,

adult females will respond to the administration of and synaptic changes. There is a neural reorganiza-
testosterone by beginning to sing and rapidly going tion that includes massive synaptogenesis associated
through the sensorimotor phase of learning. These with the incorporation of new neurons into the vocal
injections also induce marked growth of the song nu- pathways. Behavioral evidence clearly implicates
clei. The effects of testosterone in stimulating these NMDA receptor activation in specific song nuclei as
cellular changes in HVc are mediated at least in part being required for song learning (Basham, Nordeen,
via the upregulation of brain-derived neurotrophic and Nordeen, 1996). However, scientists have yet to
factor, or BDNF (Rasika, Alvarez-Buylla, and Notte- identify the precise cellular events that occur develop-
bohm, 1999). Interestingly, there is also evidence that mentally that mediate song learning rather than the
BDNF is released in HVc in response to singing itself. maturation of the song system.
Thus testosterone can promote morphological
changes in HVc by acting directly on cells in that nu-
cleus or by acting elsewhere in the brain (such as the Conclusion
preoptic area) to induce increased singing behavior Bird song is a complex motor behavior that is
that results in increased BDNF in HVc (Ball, Riters, learned by matching vocal output to an auditory
and Balthazart, 2002). memory. This memory has an innate component but
is also modified by experience early in ontogeny. In
The influence of sex steroids on the development
some species, only songs learned early in life are pro-
of the song system has been extensively studied in
duced in adulthood. In other species, learning con-
zebra finches, where the differences between the male
tinues throughout adult life. Birds tend to learn more
and female song systems are especially pronounced
songs than will be used throughout adult life, and
(see Figure 5). Female zebra finches have very small
these are selected based on social interactions among
and shrunken song nuclei and, unlike canaries, do
conspecifics in the area they settle.
not respond with song to testosterone administration
in adulthood. If given estrogen early in life, however, A specialized neural circuit has evolved in associa-
female zebra finch chicks develop masculinized song tion with the occurrence of song learning. This circuit
nuclei (Schlinger, 1998). Based on many studies of contains nuclei primarily involved in motor produc-
zebra finches, researchers are clear that this differen- tion or in the auditory feedback needed for the learn-
tiation process that occurs early in ontogeny is not the ing and maintenance of song. Neurons within this cir-
result of sex differences in gonadal steroid hormone cuit appear to be feature detectors that exhibit highly
action as is the case in the mammalian brain and in selective response to conspecific song. The selectivity
other aspects of sexually dimorphic behaviors in birds of these neurons develops in parallel with song learn-
(Schlinger, 1998). Rather, it is either the result of es- ing. The morphology of the song system varies be-
trogen synthesized by the brain acting early in ontog- tween males and females in a systematic fashion
eny to masculinize the system (Holloway and Clayton, among different species that reflects species-
2001) or of sex differences in gene expression in the variability in song behavior. Steroid hormones regu-
brain that is triggered independently of the gonad. late these sex differences as well as the striking sea-
sonal plasticity in adulthood. Songbirds exhibit adult
Another unusual feature of the songbird brain neurogenesis that contributes to the unusual adult
and avian brains in general is that neurogenesis con- plasticity. Scientists have described cellular changes
tinues to occur in portions of the periventricular zone in synaptogenesis and neuronal incorporation that
in adult birds, and these newly generated neurons mi- correlate with song learning. This system will contin-
grate out into the forebrain and are incorporated into ue to be a valuable model for the investigation of the
the neural circuitry. Some of these neurons in HVc neurobiology of species-typical learning.
project to RA (Alvarez-Buylla and Kirn, 1997). Re-
searchers are not clear, however, what significance See also: IMPRINTING; NEUROGENESIS
this phenomenon has for song learning. The new
neurons occur throughout the forebrain, in females Bibliography
as well as males. They are found in one-time learners Alvarez-Buylla, A., and Kirn, J. R. (1997). Birth, migration, incor-
such as zebra finches as well as in open-ended learn- poration, and death of vocal control neurons in adult song-
ers like canaries, and even in nonsongbirds. Nonethe- birds. Journal of Neurobiology 33, 585601.
Ball, G. F., and Hulse, S. H. (1998). Bird song. Am. Psych. 53, 37
less, this phenomenon suggests that adult birds retain
a remarkable ability to generate new neurons as well Ball, G. F., Riters, L. V., and Balthazart, J. (2002). Neuroen-
as a preservation of the cues for axon guidance and docrinology of song behavior and avian brain plasticity: Mul-
neuronal specification in their brains. tiple sites of action of sex steroid hormones. Frontiers in
Neuroendocrinology 23, 137178.
Vocal learning early in development during the Baptista, L. F., and Gaunt, S. L. L. (1997). Social interactions and
sensory phase is associated with prominent cellular vocal development in birds. In C. T. Snowdon and M. Haus-

berger, eds., Social influences on vocal development. Cambridge, McCasland, J. S. (1987). Neuronal control of bird song production.
UK: Cambridge University Press. Journal of Neuroscience 7, 2339.
Basham, M. E., Nordeen, E. J., and Nordeen, K. W. (1996). Block- Mooney, R., and Konishi, M. (1991). Two distinct inputs to an
ade of NMDA receptors in the anterior forebrain impairs sen- avian song nucleus activate different glutamate receptor sub-
sory acquisition in the zebra finch (Poephila guttata). Neuro- types on individual neurons. Proceedings of the National Acade-
biol. Learn. Mem. 66, 295304. my of Sciences of the United States of America 88, 4,0754,079.
Bottjer, S. W., and Johnson, F. (1997). Circuits, hormones, and Nelson, D. A. (1997). Social interactions and sensitive phases for
learning: Vocal behavior in songbirds. Journal of Neurobiology song learning: A critical review. In C. T. Snowdon and M.
33, 602618. Hausberger, eds., Social influences on vocal development. Cam-
bridge, UK: Cambridge University Press.
Brainard, M. S., and Doupe, A. J. (2000). Interruption of a basal
(2000). A preference for own-subspecies song guides vocal
ganglia-forebrain circuit prevents plasticity of learned vocal-
learning in a song bird. Proceedings of the National Academy of
izations. Nature 404, 762766.
Sciences of the United States of America 97 (24), 13,34813,353.
Doupe, A. J., and Kuhl, P. K. (1999). Birdsong and human speech:
Nordeen, K. W., and Nordeen, E. J. (1992). Auditory feedback in
Common themes and mechanisms. Annual Review of Neuro-
necessary for the maintenance of stereotyped song in adult
science 22, 567631.
zebra finches. Behavioral and Neural Biology 57, 5866.
Gahr, M. (2000). Neural song control system of hummingbirds:
Nottebohm, F. (1981). A brain for all seasons: Cyclical anatomical
Comparison to swifts, vocal learning (songbirds) and non- changes in song-control nuclei of the canary brain. Science
learning (suboscines) passerines, and vocal learning (budgeri- 214, 1,3681,370.
gars) and nonlearning (dove, owl, gull, quail, chicken) non- (1996). The King Solomon lectures in neuroethology: A
passerines. Journal of Comparative Neurology 426, 182196. white canary on Mount Acropolis. Journal of Comparative
Holloway, C. C., and Clayton, D. E. (2001). Estrogen synthesis in Neurophysiology A. 179, 149156.
the male brain triggers development of the avian song control Nottebohm, F., and Arnold, A. P. (1976). Sexual dimorphism in
pathway in vitro. Nature Neuroscience 4 (2), 170175. the vocal control areas in the song bird brain. Science 194,
Iyengar, S., and Bottjer, S. W. (2002). The role of auditory experi- 211213.
ence in the formation of neural circuits underlying vocal Rasika, S., Alvarez-Buylla, A., and Nottebohm, F. (1999). BDNF
learning in zebra finches. Journal of Neuroscience 22 (3), 946 mediates the effects of testosterone on the survival of new
958. neurons in an adult brain. Neuron 22, 5362.
Janata, P, and Margoliash, D. (1999). Gradual emergence of song Schlinger, B. A. (1998). Sexual differentiation of avian brain and
selectivity in sensorimotor structures of the male zebra finch behavior: Current views on gonadal hormone-dependent and
song system. Journal of Neuroscience 19, 5,1085,118. independent mechanisms. Annual Review of Physiology 60,
Jarvis, E. D., and Mello, C. V. (2000). Molecular mapping of brain 407429.
areas involved in parrot vocal communication. Journal of Com- Solis, M. M., and Doupe, A. J. (2000). Compromised neural selec-
parative Neurology 419, 131. tivity for song in birds with impaired sensorimotor learning.
Jarvis, E. D., and Nottebohm, F. (1997). Motor-driven gene ex- Neuron 25 (1), 109121.
pression. Proceedings of the National Academy of Sciences of the Suthers, R. A., Goller, F., and Pytte, C. (1999). The neuromuscular
United States of America 94, 4,0974,102. control of birdsong. Philosophical Transactions of the Royal Soci-
Jarvis, E. D., Ribeiro, S., Da Silva, M. L., Ventura, D., Vielliard, J., ety of London [Biol.] 354, 927939.
and Mello, C. V. (2000). Behaviourally driven gene expres- Tramontin, A. D., and Brenowitz, E. A. (2000). Seasonal plasticity
sion reveals song nuclei in hummingbird brain. Nature 406, in the adult brain. TINS 23, 251258.
628632. Vates, G. E., Broome, B. M., Mello, C. V., and Nottebohm, F.
Konishi, M., and Akutagawa, E. (1985). Neuronal growth, atrophy (1996). Auditory pathways of caudal telencephalon and their
and death in a sexually dimorphic song nucleus in the zebra relation to the song system of adult male zebra finches
finch brain. Nature 315, 145147. (Taenopygia guttata). Journal of Comparative Neurology 366,
Leonardo, A., and Konishi, M. (1999). Decrystallization of adult 613642.
birdsong by perturbation of auditory feedback. Nature 399 Vicario, D. S. (1991). Organization of the zebra finch song control
(6,735), 466470. system II: Functional organization of outputs from nucleus
Robustus archistriatalis. Journal of Comparative Neurology 309,
Livingston, F. S., White, S. A., and Mooney, R. (2000). Slow
NMDA-EPSCs at synapses critical for song development are
Wild, J. M. (1997). Neural pathways for the control of birdsong
not required for song learning in zebra finches. Nature Neuro-
production. Journal of Neurobiology 33, 653670.
science 3,482488.
Yu, A. C., and Margoliash, D. (1996). Temporal hierarchical con-
MacDougall-Shackleton, E. A., and MacDougall-Shackleton, S. A.
trol of singing in birds. Science 273, 1,8711,875.
(2001). Cultural and genetic evolution in mountain white-
crowned sparrows: Song dialects are associated with popula- Peter R. Marler
tion structure. Evolution 55 (12), 2,5682,575. Allison J. Doupe
MacDougall-Shackleton, S. A., and Ball, G. F. (1999). Comparative
Revised by Gregory F. Ball and Jacques Balthazart
studies of sex differences in the song-control system of song-
birds. Trends in Neuroscience 22, 432436.
Margoliash, D. (1986). Preference for autogenous song by auditory
neurons in a song system nucleus of the white-crowned spar-
row. Journal of Neuroscience 6, 1,6431,661. BLOCKING
Marler, P. (1997). Three models of song learning: Evidence from
behavior. Journal of Neurobiology 33, 501516. See: CONDITIONING, CELLULAR AND NETWORK
Marler, P., and Nelson, D. A. (1993). Action-based learning: A new SCHEMES FOR HIGHER-ORDER FEATURES OF;
form of developmental plasticity in bird song. Netherlands KAMINS BLOCKING EFFECT: NEURONAL
Journal of Zoology 43 (12), 91103. SUBSTRATES
CAJAL, RAMN Y SANTIAGO How Memory Develops During Childhood
See: RAMN Y CAJAL, SANTIAGO and Adolescence
Two factors play a critical role in the develop-
ment of memory during childhood and adolescence:
greater use of strategies and greater task-relevant
Developmental Change in Use of Memory
A strategy is any deliberate act that is designed to
CEREBELLUM improve retention. One common strategy is rehears-
See: GUIDE TO THE ANATOMY OF THE BRAIN; al, in which people spontaneously repeat (either
NEURAL COMPUTATION aloud or silently) information to be remembered. As
children grow they are more likely to use strategies to
help them remember. Rehearsal, for example, typi-
CEREBRAL NEOCORTEX cally emerges at about six or seven years of age. In ad-
dition, as children and adolescents develop, they tend
See: GUIDE TO THE ANATOMY OF THE BRAIN to abandon simple but relatively ineffective strategies
in favor of complicated but more effective strategies.
In the case of rehearsal, school-age children often
simply repeat the information exactly as it was pres-
ented. In contrast, older children and adolescents are
more likely to modify their rehearsal as necessary to
CHILDREN, DEVELOPMENT OF fit the nature of the material. Instead of repeating
MEMORY IN words in the order presented, older children and ado-
Memory improves dramatically as children develop. lescents may reorganize the words, rehearsing togeth-
The first section of this entry describes the factors that er words that are related semantically.
contribute to improved memory during childhood The general trend for children to use strategies
and adolescence. With this general developmental more often and more effectively as they grow appar-
profile as a backdrop, several special topics in chil- ently reflects parallel developmental changes in chil-
drens memory are discussed in the remainder of the drens ability to diagnose memory tasks and to moni-
entry. tor the effectiveness of their chosen strategy. That is,


Figure 1 that they are more likely to be able to provide a code

that reduces the amount of material that must be re-
Another way that childrens growing knowledge
benefits memory is by providing more retrieval cues.
Each of the links in Figure 1 provides an alternative
way, a cue, to gain access to a word that was presented.
Thus, the more links the better. When people are ex-
pert in some domain, their knowledge will be linked
extensively. If they know little, their knowledge will
have relatively few links. Consequently, older chil-
dren and adolescents will typically have more cues
available to recall information than will younger chil-
Knowledge is not always beneficial, because it can
distort memories. Information to be remembered is
often changed so that it conforms to ones existing
knowledge or stereotypes. To illustrate, when chil-
dren hear stories in which superheroes are said to be
weak or ugly, they tend to recall the heroes as strong
and attractive. Similarly, if asked to remember the ac-
tions of boys and girls whose behaviors violate sex-
role stereotypes (e.g., a girl sawing wood), children
will change the sex of the actor to make it consistent
A representation of a portion of a persons knowledge of with sex-role knowledge.

Special Topics
young children often underestimate the difficulty of The remainder of this entry considers three spe-
memory problems, and consequently may underesti- cial topics: origins of memory, childrens eyewitness
mate the need for a strategy to remember effectively. testimony, and working memory.
Also, young children are less likely than older chil-
dren to realize that a strategy is not working and Origins and Early Development of Memory
should be abandoned for a better strategy (Kail, Most methods used to study memory in children
1990). and adolescents cannot be used with infants because
they require speaking or writing. Consequently, the
Knowledge and Memory study of infants memory became possible only after
As children develop, they acquire more and more psychologists devised techniques that took advantage
knowledge of their worlds. Such understanding is an of infants abilities to respond in other ways. One ap-
important aid to memory. To explain the impact of proach is to show a stimulus to an infant; immediately
knowledge on memory, psychologists represent a thereafter, the infant is shown the same stimulus with
persons knowledge as a network like the one shown a novel stimulus. Infants typically look longer at the
in Figure 1. This network consists of nodes (the ellip- novel stimulus, behavior possible only if they remem-
ses) linked by different types of associations: isa links ber the one stimulus from its original presentation.
denote category membership; can and has denote Newborns tested on novelty-preference tasks look
properties of the node. longer at novel stimuli, indicating that some form of
memory is functional at birth.
Knowledge like that depicted in Figure 1 can aid
memory because it provides special codes that simpli- In another method used to assess infants memo-
fy memorization. To illustrate, suppose that people ry, a mobile is placed over an infants crib and a rib-
were asked to remember a list of words like collie, terri- bon connects the infants leg to the mobile. Infants
er, Dalmatian, poodle. In this case, the category name kick vigorously just a few minutes after their leg is
dogs serves as a code for the list. People can recall the connected to the mobile, demonstrating that they
category name, then scan memory for words associat- have learned the relation between their kicking and
ed with that name and decide if they were presented. the mobiles movement. To study infant memory,
Thus, one benefit of childrens growing knowledge is time is allowed to pass, then the infants leg is recon-

nected to the mobile. If the infant immediately begins age. Third, with development, children can recall
to kick vigorously, then the infant apparently has re- events over increasingly large intervals, from four
membered the relation between his or her kicking weeks at thirteen months of age to one year at twenty
and the mobiles movement. If, instead, kicking grad- months of age. Fourth, infants and toddlers recall
ually becomes more vigorousas if the infant is re- more accurately when they experience events more
learning the relationthen the infant has forgotten than once and when they actively participate in the
the relation. events (e.g., imitate the actions once before the
Research based on techniques like these has dem-
onstrated that by the age of three months infants will Childrens Eyewitness Testimony
remember the link between kicking and the moving Memory for past events is particularly important
mobile for up to fourteen days. That is, when an in- when children are asked to provide eyewitness testi-
fants leg is reconnected to the mobile within fourteen mony, as they sometimes are in cases of child abuse
days of the original learning, the infant will kick vig- (Bruck and Ceci, 1999). Although children, even
orously, without the need to relearn the relation. three- to five-year-olds, can remember past events ac-
When a sufficiently long interval had elapsed such curately, many commonly used legal procedures can
that infants no longer kick (i.e., they apparently have lead children to confuse what actually happened with
forgotten that kicking moves the mobile), a cue helps what others suggest may have happened. This is espe-
infant to remember. If the experimenter moves the cially true when interviewers are biasedthat is, when
mobile for the infant, who is later connected to the interviewers believe they know what happened and
mobile with the ribbon, the infant will again kick, ap- use questioning to attempt to confirm these beliefs.
parently signifying that the infant remembers the link The typical result is a highly suggestive interviewing
between kicking and the moving mobiles. Thus, ex- technique that is unlikely to help children recall
periments show that three important features of events accurately. For example, interviewers often ask
memory exist as early as two and three months of age: misleading questions (e.g., ones that imply events
an event from the past is remembered; over time, the happened when they may not have happened) or ask
event can no longer be recalled; and a cue can serve the same question repeatedly (implying that the
to dredge up a memory that seems to have been for- childs previous answers were wrong); both proce-
gotten (Rovee-Collier, 1999). dures increase the chances that a child will describe
Both of these memory tasks demonstrate that in- events that never happened.
fants are able to recognize events experienced previ- One particularly controversial practice is the use
ously. Other methods are used to assess recall, which of anatomically correct dolls to aid childrens recall.
refers to retrieving from memory a representation of Many professionals are strong advocates of these
a past experience that is not currently perceptible dolls, claiming that they improve memory by over-
(e.g., remembering what one had for dinner yester- coming childrens embarrassment over talking about
day). With older children and adults, recall is usually private events or by overcoming their ignorance of
assessed by having people describe verbally what hap- the names of body parts. In fact, research indicates
pened in the past. This method is not appropriate for that the use of anatomically correct dolls increases the
infants and toddlers with limited language; instead, number of false reports, apparently because the ex-
researchers have relied upon imitation tasks. An ex- aminers description of the dolls implicitly encour-
perimenter demonstrates novel actions with simple ages children to create stories about body parts.
toys; the toys are then presented to childreneither
immediately or after a delayand children are en- To improve the reliability of childrens testimony,
couraged to play with the toys. Accurate imitation of they should not be questioned repeatedly on a single
the actions is taken as evidence of recall memory, par- issue and they should be warned that interviewers
ticularly when children reproduce multiple actions in may sometimes try to trick them (e.g., suggest things
the correct order. that didnt happen). In addition, interviewers should
be neutral in their questioning and evaluate alterna-
Research with this method has revealed several tive explanations of a particular event and who was in-
important features of early recall memory. First, re- volved.
call emerges at about nine months: At this age infants
can imitate novel events after seeing them modeled Working Memory
once. Second, the amount of information that chil- Working memory denotes a cognitive structure
dren can recall immediatelydefined as the number that includes ongoing information processing as well
of actions from the sequence that are reproduced ac- as the data required for those processes. According to
curatelyincreases steadily from two actions at elev- one widely accepted view (Baddeley, 1992), working
en months of age to eight actions at thirty months of memory includes a central executive as well as two
74 CLASSICAL CONDITIONING: Behavioral Phenomena

separate subsystems for storing verbal and spatial in- EXPERIENCE AND LEARNING; FALSE MEMORIES;
formation. To measure working memory, investiga- INFANCY, MEMORY IN; WORKING MEMORY:
tors often ask people to remember information while HUMANS
concurrently performing other processing tasks. Par-
ticipants may be asked to read sets of sentences and
Baddeley, A. (1992). Working memory. Science 255, 556559.
concurrently remember the last word from each sen- Bruck, M. B., and Ceci, S. J. (1999). The suggestibility of childrens
tence in the set. Testing begins with one or two sen- memory. Annual Review of Psychology 50, 419439.
tences (hence, participants must remember one or Kail, R. (1990). The development of memory in children, 3rd edition.
two sentence-ending words) and continues until par- New York: W. H. Freeman.
(2002). Information processing and memory. In M. Born-
ticipants are no longer able to recall the sentence-
stein, L. Davidson, C. L. M. Keyes, K. A. Moore, and The Cen-
ending words in the correct sequence. Assessed in this ter for Child Well-Being, eds., Well-being: Positive development
manner, working memory increases substantially dur- across the life course. Mahwah, NJ: Erlbaum.
ing childhood and adolescence. Rovee-Collier, C. (1999). The development of infant memory. Cur-
rent Directions in Psychological Science 8, 8085.
Researchers have identified the causes and conse-
quences of age-related change in working memory. Robert V. Kail
Regarding the causes, age-related changes in working Revised by Robert V. Kail and Meghan Saweikis
memory are due primarily to age-related increases in
the speed with which children can execute basic cog-
nitive processes. As children develop, their process-
ing speed increases, which allows them to execute the
various updating functions of working memory more CLASSICAL CONDITIONING:
rapidly. For example, more rapid processing speed BEHAVIORAL PHENOMENA
results in more rapid rehearsal and more rapid re- Classical conditioning involves learning the relations
trieval of items in working memory. between stimuli. In its simplest form, a neutral stimu-
Regarding the consequences, age-related change lus precedes a stimulus (the unconditioned stimulus,
in working memory contributes to improved reason- or US) that elicits a response (the unconditioned re-
ing and problem solving during infancy, childhood, sponse, or UR). Learning is indexed by the develop-
and adolescence. That is, in the course of reasoning ment of a response (the conditioned response, or CR)
and solving problems, individuals must remember to the neutral stimulus (which is now a conditioned
task elements and coordinate task-relevant opera- stimulus, or CS). The interval between the onset of
tions, which are functions attributable to working the CS and the onset of the US is called the intersti-
memory. Consequently, as the capacity of working mulus interval (ISI). Stimuli that can become CSs may
memory increases with age, children have more re- be discrete or more contextual, and they need not
sources available to storing and processing operations even be external (Bouton, Mineka, and Barlow,
during reasoning and problem solving, resulting in 2001). Responses to stimuli (both CRs and URs) may
improved performance (Kail, 2002). Thus, as chil- be as simple as an eye blink or more complex, such
dren develop, they process information more rapidly, as approach and withdrawal. Originally thought to be
which allows them to use working memory more ef- due simply to contiguity between the CS and US,
fectively. This, in turn, produces improved reasoning modern conceptions of learning in classical condi-
and problem solving. tioning emphasize that the CS must provide informa-
tion about the US, and that the CR is both elicited by
the CS and anticipates the US.
Memory improves during childhood and adoles- Excitatory Classical Conditioning
cence, due to age-related increases in use of memory
In excitatory procedures, the CS signals that a US
strategies and age-related increases in knowledge.
will follow. Learning is indexed by the probability,
Recognition and recall are both evident in the first
magnitude, and latency of CRs to the CS. In the delay
year of life. Young children can recall past events ac-
procedure, CS onset precedes US onset and the CS
curately during legal proceedings, but biased inter-
remains on when the US is delivered. In the trace pro-
viewers may alter childrens testimony. Working
cedure, CS onset precedes US onset but the CS termi-
memory increases with age and contributes to age-
nates before the US is delivered, leaving an empty
related improvement in reasoning and problem solv-
trace interval between CS and US. Evidence sug-
gests that CS offset, as well as the CS itself, may elicit
See also: AMNESIA, INFANTILE; CODING PROCESSES: CRs in the trace procedure. Kehoe and Weidemann
ORGANIZATION OF MEMORY; EARLY (1999) tested rabbits in eyeblink conditioning using
CLASSICAL CONDITIONING: Behavioral Phenomena 75

a tone CS and a corneal air puff US in a trace proce- US (CS1-US trials) and becomes a conditioned exci-
dure. Typically, longer CS-US intervals slow condi- tor. On other trials, the two CSs are presented simul-
tioning relative to shorter CS-US intervals. However, taneously (CS1/CS2-alone trials) and no US is pres-
across CS-US intervals of 400 milliseconds to thirty ented. The CS that is never paired with the US (CS2)
seconds, the learning rate was similar, as long as the becomes a conditioned inhibitor. Another common
interval between CS offset and the US was held cons- procedure for producing inhibitory CSs is the dis-
tant at 400 milliseconds. crimination procedure, in which one CS (CS+) is fol-
lowed by the US and the other CS (CS) is not. In this
Another procedure for producing excitatory con- case, CS+ becomes excitatory and CS becomes in-
ditioning is the discrimination procedure. In simple hibitory.
discrimination conditioning, two CSs that differ along
some dimension are used. One CS is always followed Because conditioned inhibition involves the sup-
by the US (CS+) whereas the other CS is never fol- pression of CRs, which renders them unobservable,
lowed by the US (CS). Typically, CRs develop to both special tests have been developed to detect condi-
CSs initially, and then decline to the CS. This proce- tioned inhibition (Rescorla, 1969). In the retardation
dure is thought to produce both an excitatory CS test, the potential inhibitory CS, when paired with a
(CS+) and an inhibitory CS (CS, discussed further US, must take longer to elicit CRs than a new CS in
below). A more complicated type of discrimination is order to be considered a conditioned inhibitor. In the
occasion setting. In occasion setting, one CS is used. An- summation test, the potential inhibitory CS is presented
other stimulus (the occasion setter) signals whether simultaneously with an excitatory CS. If fewer CRs are
the CS will be paired or unpaired on a particular trial. observed to this compound than to the excitatory CS
This occasion setter is not considered a CS because it alone, the test CS is considered a conditioned inhibi-
does not appear to form a direct association with the tor.
US. An example of this procedure, in which a contex-
tual stimulus served as the occasion setter, is provided Unlike the widespread acceptance of excitatory
by Rogers and Steinmetz (1998). In that study, a flash- associations between a CS and US (but see Gallistel
ing chamber light was the occasion setter. Whenever and Gibbon, 2000, for an alternative), inhibitory asso-
the flashing chamber light was on, the context was lit ciations have been controversial (Savastano, Cole,
and a tone CS was followed by a corneal air puff US. Barnet, and Miller, 1999). The most common view
Whenever the flashing chamber light was off, the con- has been that inhibitory associations are simply the
text was darkened, and the same tone CS was not fol- opposite of excitatory associations (Rescorla and
lowed by the US. Rabbits learned CRs when the Wagner, 1972). In this view, a single continuum for
chamber was lit and suppressed CRs when the cham- association exists, ranging from positive (excitatory)
ber was darkened. The light or dark chamber by itself to negative (inhibitory). Other researchers have pro-
cannot elicit CRs, so it is not a CS. Rather, the state posed two separate continuums, one for excitatory as-
of the context served to modulate the development of sociations and the other for inhibitory associations
CRs to the tone CS. (Pearce and Hall, 1980). These two continuums op-
pose each other and the presence or absence of CRs
is a reflection of this opposition. Finally, Miller and
Inhibitory Classical Conditioning Matzel (1988) have developed a model of classical
conditioning in which inhibitory associations do not
In inhibitory procedures, the CS signals that a US exist at all, and the presence or suppression of CRs
will not follow. It is important to note that an inhibito- is determined by the inequality of excitatory associa-
ry CS, like an excitatory CS, provides information tions for different stimuli.
about the US. Thus, a US must be present somewhere
in the conditioning situation for inhibitory condition-
ing to occur. This is in contrast to latent inhibition Stimulus Preexposure
(discussed under Stimulus Preexposure, below).
The fact that a US must be present somewhere in the Exposure to the stimuli by themselves prior to
conditioning situation means that some other stimu- classical conditioning can change the rate at which
lus in the conditioning situation, either discrete or these stimuli become CSs. The most well-studied ex-
contextual, becomes a conditioned excitor. ample of stimulus preexposure involves preexposure
to the CS, or latent inhibition (Lubow and Moore,
A number of procedures are thought to be able 1959). In contrast to a new CS, a preexposed CS that
to produce inhibitory CSs. The most common proce- is subsequently paired with a US will come to elicit
dure was originally used by Ivan Pavlov (1927). In CRs only slowly. Because no US is present during pre-
Pavlovs conditioned inhibition procedure, two CSs are exposure, the CS does not provide information on ei-
used. On some trials, one of the CSs is paired with a ther the presence or absence of the US during pre-
76 CLASSICAL CONDITIONING: Behavioral Phenomena

exposure and, therefore, forms neither an excitatory text-US association, showed less blocking than rabbits
nor an inhibitory association. Rather, the preexposed that did not receive context-alone exposure. In addi-
CS is thought to undergo a loss of salience, and is tion, rabbits shifted to a different context between
treated as irrelevant. phases 1 and 2 also showed less blocking than rabbits
Schmajuk, Lam, and Gray (1996) developed a de- training in the same context in phases 1 and 2.
tailed neural network model in which latent inhibi-
tion is explained by a reduction in the ability to store
CR Timing
and retrieve CS-US associations. Preexposure to a CS
reduces both attention to and the magnitude of the In classical conditioning, a US elicits a UR and a
representation of that CS. During subsequent CS-US CS comes to elicit a CR. However, the CR is not just
pairings, less attention to the CS slows the storage of elicited by the CS. In addition, the CR is made in an-
an excitatory CS-US association and a less intense ticipation of the US. In many classical conditioning
representation slows the ability of the CS to elicit re- procedures, such as eyeblink conditioning, this antici-
trieval of the CS-US association. pation takes the form of precise timing of the CR to
occur just before the onset of the US.

Cue Competition There are several ways to demonstrate timing of

CRs. The simplest method is ISI shift. In an ISI-shift
Multiple stimuli may be paired with a US, but not
procedure, training initially occurs at one CS-US in-
all of them will develop the ability to elicit a CR. Con-
terval. When the CS-US interval is lengthened or
sider two CSs presented simultaneously and followed
shortened, the latency of CRs to the CS lengthens or
by a US. If the CSs are similar in salience, each of the
shortens to coincide with the new CS-US interval. An-
CSs will acquire the ability to elicit a CR when pres-
other procedure involves a dual ISI. In a dual-ISI pro-
ented separately. However, if the CSs differ greatly in
cedure, the CS-US interval varies between two values
salience (for example, a bright light and a soft tone),
across trials. Typically, this procedure produces a
only one of the two CSs will come to elicit CRs. The
double-peaked CR, with the peaks corresponding to
more salient CS is said to overshadow, or disrupt, the
the two CS-US intervals. Finally, timing of CRs can be
formation of an association between the less salient
demonstrated with ISI discrimination. In an ISI-
CS and the US. This phenomenon demonstrates that
discrimination procedure, two CSs that differ along
mere contiguity of a CS and a US is not always enough
some dimension are used, as in a standard discrimi-
to produce an association between them.
nation procedure. However, in contrast to a standard
The insufficiency of contiguity between a CS and discrimination procedure, each CS is paired with a
a US for learning in classical conditioning was force- US, but at a different CS-US interval. Mauk and Ruiz
fully demonstrated by Kamin (1969). Kamins proce- (1992) used tones of different frequencies for the two
dure, which is known as blocking, involves three CSs and a corneal air puff as the US. Rabbits learned
phases. In phase 1, a CS is paired with a US (CS1-US a CR to each CS with a latency appropriate to the
trials) until it becomes a strong conditioned excitor. CS-US interval.
In phase 2, the excitatory CS is presented in com-
pound with a new CS, and the entire compound is
paired with a US (CS1/CS2-US trials). In phase 3, the Conclusion
two CSs are tested separately (CS1-alone and CS2- Although classical conditioning involves learning
alone trials) for the ability to elicit a CR. The typical the relations between two stimuli, these relations can
result is that CRs occur only to the CS used in phase be complex. The rate and content of learning in clas-
1 (CS1). This demonstrates that a CS must provide sical conditioning depends upon a number of factors.
new information about the US in order to form an as- These include the following: 1. whether the CS is fol-
sociation with it. Mere contiguity between a CS and lowed by a US or not; 2. how many CSs are present;
a US is not sufficient for classical conditioning. 3. whether the organism has any previous experience
Contextual cues may contribute to the blocking with the CS, the US, or the relationship between the
effect. In essence, it has been proposed that the CS two; and 4. how much time elapses between the deliv-
and the context in phase 1 form a combined associa- ery of the CS and the delivery of US. In addition, the
tion with the US and this combined association blocks stimuli used as CSs can be as complex as an entire
learning to the other CS in phase 2. Giftakis and Tait context. Although the responses examined in labora-
(1998) tested rabbits using a tone CS, a flashing light tory studies of classical conditioning are often simple
CS, and a periorbital shock US. Rabbits given several (such as an eye blink), this is simply a control and
sessions of exposure to the context alone between measurement issue. Classical conditioning can in-
phases 1 and 2, which serves to extinguish any con- volve complex learned responses as well, and is be-

lieved to play an important role in everyday human CODING PROCESSES

[Coding processes refers to the ways in which information
may be represented in memory. Events in the world strike our
See also: CONDITIONING, CELLULAR AND NETWORK senses and may be well perceived, but the mental operations
that ensue determine whether the events will be remembered.
These mental operations are referred to as coding processes
GENERALIZATION; KAMINS BLOCKING EFFECT: and have been studied in several different ways. Three types
NEURONAL SUBSTRATES; NEURAL SUBSTRATES of coding processes are discussed in the entries in this section.
OF CLASSICAL CONDITIONING: DISCRETE The first entry is on coding processes that involve mental
BEHAVIORAL RESPONSES IMAGERY. People tend to remember information better if they
convert the information to mental pictures while they study
it. This technique is frequently used by experts who can mem-
Bibliography orize huge amounts of information (see MNEMONISTS) and
Bouton, M. E., Mineka, S., and Barlow, D. H. (2001). A modern
by all people who employ effective memory strategies (see
learning theory perspective on the etiology of panic disorder.
Psychological Review 108, 432. MNEMONIC DEVICES). In the LEVELS OF PROCESSING
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition- approach, people are directed to think about different aspects
ing. Psychological Review 107, 289344. of events (attention is directed to superficial properties of
Giftakis, J. E., and Tait, R. W. (1998). Blocking of the rabbits classi- events or their meanings) and memory is tested later. The
cally conditioned nictitating membrane response: Effects of
deeper or more meaningful the level of processing, the better
modifications of contextual associative strength. Learning and
Motivation 29, 2348. the later memory under most circumstances. The final topic
Kamin, L. J. (1969). Predictability, surprise, attention, and condi- in this section is more general, about ORGANIZATION OF
tioning. In B. A. Campbell and R. M. Church, eds., Punishment MEMORY. One effective strategy to code information is to or-
and aversive behavior, pp. 279296. New York: Appleton- ganize it in terms of knowledge we already have. The study
of memory organization is concerned with both how knowl-
Kehoe, E. J., and Weidemann, G. (1999). Within-stimulus competi-
tion in trace conditioning of the rabbits nictitating mem- edge is organized and how people use their knowledge to en-
brane response. Psychobiology 27, 7284. code new information in memory. The study of coding pro-
Lubow, R. E., and Moore, A. U. (1959). Latent inhibition: The ef- cesses is central to the study of human memory and ramifies
fect of nonreinforced preexposure to the conditioned stimu- through most other topics. For example, whether some bit of
lus. Journal of Comparative and Physiological Psychology 52, 415
information can be retrieved from memory depends on how
Mauk, M. D., and Ruiz, B. P. (1992). Learning-dependent timing it was encoded when it was learned (see RETRIEVAL PRO-
of Pavlovian eyelid responses: Differential conditioning using CESSES IN MEMORY).]
multiple interstimulus intervals. Behavioral Neuroscience 106,
Miller, R. R., and Matzel, L. D. (1988). The comparator hypothesis:
A response rule for the expression of associations. In G. H. IMAGERY
Bower, ed., The psychology of learning and motivation, pp. 5192.
San Diego, CA: Academic Press. According to Cicero, it was the Greek poet Simonides
Pavlov, I. P. (1927). Conditioned reflexes. New York: Dover. who first recognized the utility of mental imagery for
Pearce, J. M., and Hall, G. (1980). A model for Pavlovian learning: memory. During a brief absence from a banquet at
Variations in the effectiveness of conditioned but not of un-
conditioned stimuli. Psychological Review 87, 532552.
which the poet recited the entertainment, the roof of
Rescorla, R. A. (1969). Pavlovian conditioned inhibition. Psychologi- the building caved in, mangling the guests so badly
cal Bulletin 72, 7794. that recognition of the bodies was impossible. Simoni-
Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian des (who likely already had developed a fairly good
conditioning: Variations in the effectiveness of reinforcement memory in the context of his job) realized that he
and nonreinforcement. In A. H. Black and W. F. Prokasy,
eds., Classical conditioning II: Current research and theory, pp.
could remember the faces and clothing of the guests
6499. New York: Appleton-Century-Crofts. in various locations around the room; thus he was
Rogers, R. F., and Steinmetz, J. E. (1998). Contextually based con- able to identify the corpses for their families. Second-
ditional discrimination of the rabbit eyeblink response. Neuro- arily, perhaps, the art of memory was born (see Yates,
biology of Learning and Memory 69, 307319. 1966, for full histories).
Savastano, H. I., Cole, R. P., Barnet, R. C., and Miller, R. R. (1999).
Reconsidering conditioned inhibition. Learning and Motiva- Image and imagery generally are used to refer to
tion 30, 101127. those concrete, perceptual, and usually visual modes
Schmajuk, N. A., Lam, Y-W, and Gray, J. A. (1996). Latent inhibi- of thought that appear to represent the physical
tion: A neural network approach. Journal of Experimental Psy-
chology: Animal Behavior Processes 22, 321349.
world relatively directly. These are clearly distin-
guishable from verbal thought processes, which are
John T. Green arbitrary in the linguistic sense of there being no nec-
Joseph E. Steinmetz essary relationship (other than social agreement) be-

Figure 1 required for subjects to scan from one location to an-

other on their images is a direct function of the linear
distance between the two locations on the original
map, thus reflecting the analogue character of mental
images. Further studies by Kosslyn, Finke, and others
have shown that people can mentally construct two-
and three-dimensional figures that have emergent
properties not predictable from their component
parts. Such findings indicate a perceptionlike quality
of mental imagery that has been supported by some
neurophysiological evidence (Farah, 1984).
Although the psychological processes underlying
these phenomena may not be entirely clear, the utility
of mental imagery and pictures to facilitate memory
has been recognized for centuries. Simonidess method
of loci, for example, in which successive items are im-
aged at specific locations along a familiar route, has
long been used for learning ordered lists of unrelated
items or remembering a series of points to be made
during a speech. Imagery also can be helpful in learn-
ing foreign vocabulary, as in forming an image of a
cadaver falling from a table for the Italian cadere (to
Schematic diagram of the dual coding model and one possible fall) or someone harvesting cogs in a field for cogliere
alternative. (to pick or gather). In addition, imagery seems cen-
tral to the development of thinking and memory in
children, and a progression from motor processes to
tween words and their referents. Mental images of visual images to verbal processes is assumed by most
things experienced thus often appear to be funda- major theories of cognitive development.
mentally related to their meanings. It is difficult to
In scientific studies of memory, there are several
think of a pizza without some olfactory or gustatory
typical findings that indicate an important role of im-
image or to think of ones mother without experienc-
agery: Words rated as referring to more concrete,
ing a visual image of her face.
highly imageable things are better remembered than
Consistent with subjective impressions, a variety words referring to abstract, nonimageable things
of studies have indicated that visual imagery and visu- the concreteness effect; the use of imagery in learning
al perception have some similar qualities. Several (either by instruction or spontaneously) leads to bet-
studies by Shepard and his colleagues, for example, ter memory than nonimaginal strategiesthe imagery
have examined mental rotation (Cooper and Shepard, effect; and pictures are better remembered than
1973). This paradigm typically involves asking people wordsthe picture superiority effect.
to judge whether two visually presented stimuli (e.g.,
This apparent centrality of mental imagery in
letters or three-dimensional shapes) are identical or
thinking and memory has led to several theoretical
mirror reflections of each other. Shepards results
frameworks for studying human memory and cogni-
have consistently indicated that, in order to decide on
tion that give images an essential role. The most com-
the possible identity of the two stimuli, subjects first
pletely articulated of these, the dual coding model of
mentally rotate one of them to the same orientation
Paivio (1971, 1986), is depicted in Figure 1(a). The
as the other. This is evidenced by the fact that reac-
model includes a verbal system specialized for dealing
tion times to make the identity judgments increase
with sequential, especially linguistic information and
regularly as the angular difference between the two
an imagery system specialized for dealing with non-
stimuli increases from 0 to 180 degrees and then de-
verbal, holistic information. It explicitly entails that
creases to 360 degrees (as the rotation can be made
concrete experiences generate perceptual memory
traces (images) that preserve the structural attributes
Kosslyn and his colleagues have demonstrated of the input. Those generated memory representa-
similar findings in a series of studies on mental scan- tions are assumed to be functionally equivalent to per-
ning (Kosslyn, 1973). In this paradigm, subjects mem- ceptual representations in the sense that an image
orize several landmarks on a simple map and then are of an object generated to its name has the same mne-
asked to form an image of it. Consistently, the time monic properties as the image evoked by the object

itself (Paivio, 1986, p. 144). Among other conse- One difficulty in explaining the effects of imagea-
quences, the storage of information-rich perceptual bility on memory is that several other variables co-
images is assumed to account for the findings that occur with it. Dimensions such as concreteness and
mental comparisons of things on physical dimensions word frequency have been shown to be positively re-
(such as size) yield similar patterns of response times lated to rated imagery values, whereas others, such as
regardless of whether the objects, their pictures, or generality of reference and associative set size, have
their printed names are presented. been shown to be negatively related to it. The effect
of imagery can be empirically or statistically separat-
The hypothesized existence of separate but inter- ed from all of those other variables, however, and re-
connected verbal and imaginal systems in the dual mains a significant predictor of memory when the
coding model explains the concreteness effect in others are controlled (Paivio, 1986). Exceptions begin
terms of coding redundancy: Concrete words usually re- to occur only when memory for more complex mate-
sult in dual verbal and imaginal memory codes (via rials like concrete and abstract texts is considered or
the intersystem, referential crossover), whereas ab- when other meaningful contexts make the impor-
stract words usually result in only a single, verbal tance of imaginal processing less essential (Marschark
code. Such redundancy provides two alternative and Cornoldi, 1990).
routes to the traces for concrete words during recall The precise role of imagery in language compre-
and provides a backup if one code is forgotten. In hension and problem solving remains unclear, al-
learning of concrete word pairs (i.e., in paired-associate though it clearly plays a role in visually and spatially
learning), imagery is assumed to provide a means of oriented situations, such as route learning and chess
integrating the meanings of the two words into a sin- playing, which seem to involve analogue mental ma-
gle, imaginal unit that can be retrieved later in the nipulation. Distinguishing imagery from the nature
context of one of the words as a cue. Integration of of long-term memory codes highlights the role of in-
this sort is assumed to be unavailable for abstract dividual differences in imagery ability. A variety of
word pairs, which must be stored as two-unit verbal standardized and well-documented tests that involve
strings. Finally, pictures are assumed by the dual code tasks such as mental paper-folding, mental rotation,
model to be better remembered than words in part and mental scanning have provided evidence of large
because images may be inherently more memorable differences between individuals in the vividness,
than verbal traces but primarily because pictures are speed, and frequency of image generation. The imag-
likely to elicit verbal naming (and hence dual memory ery abilities tapped by these tests generally are not
codes) with a somewhat higher probability than words predictive of memory ability even for concrete materi-
elicit imagery. als. Nonetheless, it is clear that variability in some im-
agery abilities can have specific and marked effects in
One modified version of the dual coding model a variety of cognitive domains, including memory. In-
is depicted in Figure 1(b). In this alternative, dual ver- dividuals who have suffered head injuries, for exam-
bal and imaginal processing systems are assumed to ple, typically exhibit deficits in visual-spatial tasks
operate at a level akin to Baddeleys (1986) working such as finding hidden figures or reconstructing pre-
memory, which has both visual and verbal components. viously presented pictures. They may also fail to ex-
The dual processing systems can account for most of hibit concreteness effects in memory and be particu-
the results concerning verbal processes and image larly slow at making imaginal comparisons
manipulation and inspection. In contrast with the (Richardson, 1990). People who score high on tests of
multiple, modality-specific memory codes of the orig- image manipulation skill tend to be faster in compar-
inal dual coding model, however, this version as- ing their images on particular dimensions, although
sumes that long-term memory involves some more they may be no faster in generating those images in
generic, semantic, or propositional code common to the first place.
both concrete and abstract information (Potter, Among the more puzzling imagery findings yet to
1979). Information is retrieved from this conceptual be explained is why people who are totally, congeni-
memory and images are constructed in visual working tally blind still show apparent imagery and concrete-
memory. Imagery is still invoked to explain the bene- ness effects (DeBeni and Cornoldi, 1988). These ef-
ficial effects of imagery instructions, material con- fects appear not to be attributable to tactile
creteness, and picture presentation, but the locus of knowledge, because they are just as readily obtained
their effects on memory is placed on the distinctiveness when the to-be-remembered items are things for
of imaginally processed items within the encoding which tactile experience is unlikely (e.g., a moon, a
and retrieval contexts (Marschark et al., 1987). Inter- tiger, or a tower). Clearly, imagery is a multidimen-
item integration is assumed to be possible for abstract sional construct that has great theoretical and practi-
as well as concrete materials via conceptual relations. cal utility but no simple psychological explanation.
80 CODING PROCESSES: Levels of Processing

Bibliography such as whether the word is a synonym for a specified

Baddeley, A. (1986). Working memory. Oxford: Clarendon Press. word. The assumption is that these three types of
Cooper, L. A., and Shepard, R. N. (1973). The time required to judgments require increasingly deep levels of pro-
prepare for a rotated stimulus. Memory & Cognition 1, 246 cessing. In a subsequent memory test in which partici-
DeBeni, R., and Cornoldi, C. (1988). Imagery limitations in totally pants are asked to recall the words, deeper levels of
congenitally blind subjects. Journal of Experimental Psychology: processing are associated with higher levels of recall:
Learning, Memory, and Cognition 14, 650655. those words that required a semantic-level judgment
Farah, M. J. (1984). The neurological basis of mental imagery: A are recalled best, whereas those requiring ortho-
componential analysis. Cognition 18, 245272.
graphic processing yield the lowest recall level.
Kosslyn, S. M. (1973). Scanning visual images: Some structural im-
plications. Perception and Psychophysics 14, 9094. Craik and Lockhart (1972) proposed the general
Marschark, M., and Cornoldi, C. (1990). Imagery and verbal mem- concept of levels of processing not as a theory of
ory. In C. Cornoldi and M. A. McDaniel, eds., Imagery and cog-
memory but rather as a framework for future research
nition, pp. 133182. New York: Springer-Verlag.
Marschark, M., Richman, C. L., Yuille, J. C., and Hunt, R. R. into the relationship between coding processes and
(1987). The role of imagery in memory: On shared and dis- memory. Lockhart and Craik have written a retro-
tinctive information. Psychological Bulletin 102, 2841. spective commentary on the significance of this pro-
Paivio, A. (1971; reprint 1979). Imagery and verbal processes. Hills- posal (1990). The basic claim of levels of processing
dale, NJ: Erlbaum.
as a research framework is that a thorough under-
(1986). Mental representations: A dual coding approach. Ox-
ford: Oxford University Press. standing of remembering requires a careful analysis
Potter, M. C. (1979). Mundane symbolism: The relations among of the way in which (and the degree to which) coding
objects, names, and ideas. In N. R. Smith and M. B. Franklin, processes involve the construction of meaning. Ac-
eds., Symbolic functioning in childhood, pp. 4165. Hillsdale, NJ: cording to this view, there is no distinct coding pro-
Erlbaum. cess that can be identified as committing to memo-
Richardson, J. T. E. (1990). Imagery and the brain. In C. Cornoldi
and M. A. McDaniel, eds., Imagery and cognition, pp. 145. ry. Rather, memory codingthe memory traceis
New York: Springer-Verlag. constructed as a byproduct of the everyday mental op-
Yates, F. A. (1966). The art of memory. Chicago: University of Chica- erations we perform as we interact with our environ-
go Press. ment and attempt to understand it. One implication
Marc Marschark of this claim is that a proper account of the relation-
ship between coding processes and memory will in-
volve an understanding of the broader issues of per-
ception and comprehension.
Orienting Tasks
Processing and Recall
Experimental conditions (orienting tasks) such as
The term levels of processing was introduced by Craik those involving judgments of case, rhyme, or syn-
and Lockhart (1972) to describe the way in which the onymity are but three examples of the large number
information contained in a stimulus can be analyzed of orienting tasks that have been used in experiments.
at levels ranging from surface physical properties to Other examples are rating a words pleasantness, de-
deeper levels involving its meaning. Thus in reading ciding how many syllables it has, and whether it is
the printed word clever, the reader might process or- spoken by a male or a female voice. Indeed, we can
thographic features, such as its being in capital let- think of our everyday cognitive activity in terms of a
ters, or phonemic features, such as that it rhymes with continuous sequence of orienting tasks as our knowl-
ever, or semantic features, such as that it is a synonym edge, goals, and needs interact with the circum-
for skilled. stances of the moment. Our goals will influence our
The level of processing is a powerful determinant orientation toward a stimulusthe specific informa-
of how well an event will be remembered (Craik and tion that we selectively attend to and analyze. The
Tulving, 1975). A simple demonstration experiment particular form of the analysis will strongly influence
illustrates this point. Participants in the experiment later remembering. In listening to a conversation, for
are presented with a sequence of common words and example, our goal may be to comprehend what is
asked to make one of three possible judgments about being said, but it may also be to infer the speakers in-
each word. For some words participants are asked to telligence, mood, or intentions, or the origin of a dis-
make a judgment at the orthographic level, such as tinctive accent. When we read a restaurant menu, our
whether the word is printed in capital letters; other goal may involve judgments of taste, preference, cost,
words require a phonemic-level judgment, such as or nutritional value. The fundamental principle of le-
whether the word rhymes with a certain word; a third vels of processing as a research framework is the claim
set of words requires a judgment involving meaning, that since the memory trace is the byproduct of these
CODING PROCESSES: Levels of Processing 81

analyses, the key to predicting subsequent memory Skills and the Material to Be Remembered
performance is a matter of gaining increased under- Another determinant of the level of processing is
standing of the nature and level of these analyses. the nature of the material to be remembered. Pictures
Orienting tasks used in experimentstasks such as and common words afford the rapid analysis of mean-
judging rhyme or synonymity (meaning)are the sci- ing. Other material, such as proper nouns, can pose
entists effort to gain tight control over the processing greater difficulty. One reason for the common experi-
that a participant applies to a stimulus to evaluate the ence of rapidly forgetting names following introduc-
impact that different processes have on memory. tions at a party is that proper nouns do not trigger
rapid deep processing. Hence most mnemonic tech-
niques for remembering names are essentially strate-
Incidental versus Intentional Processing
gies for converting proper names into a visual image
Levels of processing have important implications that embodies deeper semantic-level processing. In-
for the distinction between incidental and intentional teracting with the nature of the material is a second
processing. Because we do not usually make an inten- determinant: the skill of the rememberer. To the
tional effort to commit everyday experiences to mem- nonspeaker of French the letter string chien is mean-
ory, much of our normal remembering is incidental. ingless, as is a musical score to anyone who has not
You can probably recall what you were doing exactly been trained to read music, or the game position of
twenty-four hours ago even though, at the time, you chess pieces to one who does not play chess. But to the
were not making a conscious effort to commit the ac- speaker of French, to the trained musician, or to the
tivity to memory. This aspect of everyday life can be skilled chess player, the word, the score, and the
captured in an experimental setting by using inciden- board position, respectively, afford deep processing
tal orienting tasks in which participants are not in- and hence better remembering.
formed that they will receive a subsequent memory
test. Participants might perform the judgments be-
lieving that the only purpose of the experiment is to Conclusion
see how quickly they can respond. According to the Levels of processing is a theoretical framework
levels of processing approach, such participants that claims that memory coding is a byproduct of cog-
should not be disadvantaged relative to those in- nitive and interpretive processes. Memory depends
structed to expect the memory test, provided the level heavily on the degree to which the processing in-
of processing for the two groups is comparable. That volves the construction and elaboration of meaning.
is to say, the important determinant of remembering Tasks such as judging case, rhyme, or synonymity
is not the conscious effort of committing something provide a clear example of three distinct levels of pro-
to memory but the level of processing that the orient- cessing. Depth of processing, however, can vary over
ing task induces. This conclusion is supported by ex- a wide range. According to levels of processing as a
perimental findings. For example, participants who research framework, research into coding processes
are asked to rate a word for pleasantness as an inci- seeks to provide a precise account of this variation
dental orienting task perform as well on a subsequent and its impact on memory. Recent developments of
unexpected memory test as those who are given prior the original idea of levels of processing and critical
warning of the test. analyses are discussed in Naveh-Benjamin, Mos-
What happens when a researcher exhorts partici- covitch, and Roediger (2001).
pants in an experiment simply to try to remember?
Presumably the subjects will process the material in Bibliography
whatever way they think will most effectively support Craik, F. I. M., and Lockhart, R. S. (1972). Levels of processing:
later remembering. A common strategy with verbal A framework for memory research. Journal of Verbal Learning
material is to rehearse it by silently repeating a word and Verbal Behavior 11, 671684.
Craik, F. I. M., and Tulving, E. (1975). Depth of processing and the
or phrase over and over. Such a strategy is relatively
retention of words in episodic memory. Journal of Experimental
ineffective for long-term remembering, since such re- Psychology: General 104, 268294.
petitive processing typically involves no further analy- Lockhart, R. S., and Craik, F. I. M. (1990). Levels of processing:
sis of meaning (no deeper-level processing) but con- A retrospective commentary on a framework for memory re-
sists of maintaining phonemic recollection. Again, the search. Canadian Journal of Psychology 44, 87112.
important point is that successful remembering is less Naveh-Benjamin, M., Moscovitch, M., and Roediger, H. L. III, eds.
(2001). Perspectives on human memory and cognitive aging: Essays
a matter of conscious effort than of being led to per- in honor of Fergus Craik. Philadelphia: Psychology Press.
form an orienting task that demands deep levels of
processing. Robert S. Lockhart
82 CODING PROCESSES: Organization of Memory

ORGANIZATION OF MEMORY at least one fundamental process of organization. As-

sociations derive from the frequent temporal cluster-
ing of events. In the early part of the twentieth centu-
Coding and Organization ry, Pavlov (1927) discovered classical conditioning.
Coding refers to the interpretations a person gives to This discovery led to extensive investigations of the
experiences. The significance of experience for mem- formation and maintenance of associations. Pavlov
ory and action depends on the interpretation of the found that after frequently presenting a neutral stim-
experience. The same events can be interpreted in ulus (e.g., a tone) in close proximity to the presenta-
dramatically different ways depending on a persons tion of food, a dog would salivate at the sound of the
knowledge and expectations. To understand coding tone even in the absence of food. Thus, an association
we must understand the organization and use of formed between the tone and the food.
knowledge in interpreting experience. The interre-
Garcia and Koelling (1966) found that some asso-
latedness of ideas is one of the most compelling facts
ciations are learned more easily than others. Their
of mental life. In personal memories, a single associa-
laboratory rats learned to associate a novel taste with
tion with some present event can trigger detailed
gastrointestinal illness much more easily than they
memories of past experiences. Psychology has devel-
learned the association between a flashing light and
oped several ideas about the nature of organization
gastrointestinal illness. This result suggests that vari-
in memory.
ous constraints influence the formation of associa-
We can illustrate the influence of coding by com- tions.
paring the memories of two people with different de-
grees of knowledge: in this case, an expert and a non-
expert about cars. They both see the same small red Associative Networks
car. The expert identifies it as a Miata; the nonexpert In the direct representation of associations in the
can identify it only as a small red car. Would it sur- form of a network, concepts are shown as nodes and
prise you if later the expert was able to state with some associations are shown by lines (or links) connecting
confidence that a small red Triumph was not the car the nodes. Schvaneveldt, Durso, and Dearholt (1989)
seen earlier, while the nonexpert had more difficulty presented a method of deriving such networks from
in making this discrimination? Each individuals proximity data such as judgments of relatedness
knowledge influences the coding and thus the memo- among sets of concepts. Cooke, Durso, and Schvane-
ry of the experience. veldt (1986) found that networks can predict the way
Human memory imposes organization on our ex- people organize the concepts when they learn a list of
periences. Tulving (1962) and others have shown that words. Goldsmith and Johnson (1990) were able to
when people learn a list of randomly selected words, predict students grades in a course on experimental
they organize the words in recalling the list. As the list methods from the degree of similarity of the students
is learned, there is more and more consistency in the and the instructors networks of important concepts.
grouping of the words in recall.
Earlier, Bousfield (1953) showed that subjects re-
Semantic Networks and Semantic Features
call lists of words as clusters of related words. For ex-
ample, if the list contained some names of flowers, Semantic networks also use network representa-
some names of people, some types of buildings, and tions, but they specify more about the relations be-
so on, then the free recall of these words would group tween concepts by using labeled links (Collins and
the similar items. This grouping occurs even though Quillian, 1969; Meyer and Schvaneveldt, 1976; Quil-
the words are presented in random order. Later lian, 1969). For example, such a network would show
Bower and his colleagues (Bower, 1970) showed that that robin is a member of the class bird with an isa
theories about the structure of memory could predict link (A robin is a bird). It would also show that a deer
the organization of material to be learned. Bransford has antlers, and so on. Such networks can also support
and Johnson (1972) studied passages that are difficult inferences such as concluding that a robin is an ani-
to remember unless people are led to give them ap- mal by retrieving a robin is a bird and a bird is an ani-
propriate interpretations. Their work is an impressive mal. Semantic networks have been used to explain ex-
demonstration of the role of interpretation in re- perimental data from studies in language
membering. understanding and category judgments. Such net-
works are also often a part of computer programs de-
signed to exhibit artificial intelligence (Quillian,
Organization of Memory 1989). Other theories propose that concepts consist
What leads to the organization of memories? of collections of features that define the concepts
Most answers to this question refer to association as (Smith and Medin, 1981). The concept bird, for exam-
CODING PROCESSES: Organization of Memory 83

ple, might consist of features such as has wings, flies, constrained by the history and situation of the indi-
lays eggs, has feathers, and so on. According to feature vidual.
theories, when people reason about concepts, they re-
trieve features from memory and use them to draw
conclusions. Conclusion
Coding is the interpretation of events in light of
what we know. Such interpretation can have benefi-
Schemata cial consequences, as in the superiority of the memory
Schemata are general representations of several of chess masters for real board positions. Sometimes
different items of information together with the speci- interpretation leads to false memories of related in-
fication of the relations among the items (Bartlett, formation that was not actually experienced (Loftus
1932; Minsky, 1975). For example, the schema for a and Ketcham, 1991). Understanding the memory of
room might specify that it must have a floor, a ceiling, an event requires an understanding of the coding that
walls, and a door as well as some spatial relations arises from cumulative knowledge. An important
among these. Optionally, it might have additional question for theory and research concerns the extent
doors and windows. Scripts are examples of schemata to which memory depends on stored representations
where actions are organized in familiar sequences as opposed to cues available from the body and the
such as going to a restaurant or visiting the doctor. environment.
Schemata invite inferences. Several studies suggest
that memory includes inferred information (defaults) See also: CODING PROCESSES: IMAGERY; CODING
in addition to what we actually experience. For exam- PROCESSES: LEVELS OF PROCESSING; FALSE
ple, if we hear the sentence, Fred drove the nail into MEMORIES
the board, we are likely to infer that he used a ham-
mer even though the sentence does not mention a Bibliography
hammer. If someone eats in a restaurant, we assume Bartlett, F. C. (1932). Remembering: A study in experimental and social
that he or she paid for the meal. psychology. Cambridge, UK: Cambridge University Press.
Bickard, M. H. (2000). Dynamic representing and representational
Chase and Simon (1973) reported a classic dem- dynamics. In E. Dietrich and A. Markman, eds., Cognitive dy-
onstration of the power of schemata using memory namics: Conceptual and representational change in humans and ma-
for the positions of pieces on a chess board. They chines. Mahwah, NJ: Erlbaum.
Bousfield, W. A. (1953). The occurrence of clustering in the recall
found that chess masters were no better than novices
of randomly arranged associates. Journal of General Psychology
at reconstructing a board with randomly placed 49, 229240.
pieces, but the masters were far superior in recalling Bower, G. H. (1970). Organizational factors in memory. Cognitive
the positions of pieces from the middle of an actual Psychology 1, 1846.
chess game. Experts presumably have elaborate sche- Bransford, J. D., and Johnson, M. K. (1972). Contextual prerequi-
sites for understanding: Some investigations of comprehen-
mata that can code the positions of the pieces on the
sion and recall. Journal of Verbal Learning and Verbal Behavior
board when the positions make sense. 11, 717726.
Chase, W. G., and Simon, H. A. (1973). Perception in chess. Cogni-
tive Psychology 4, 5581.
Embodiment and the Need for Collins, A. M., and Quillian, M. R. (1969). Retrieval time from se-
mantic memory. Journal of Verbal Learning and Verbal Behavior
Representations 8, 240247.
In recent years challenges to traditional ideas Cooke, N. M., Durso, F. T., and Schvaneveldt, R. W. (1986). Recall
and measures of memory organization. Journal of Experimental
about the role of mental representations have arisen
Psychology: Learning, Memory, and Cognition 12, 538549.
from researchers in cognitive science. A major con- Edelman, G. M. (1992). Bright air, brilliant fire: On the matter of mind.
cern is that traditional approaches have neglected the New York: Basic Books.
constraints imposed on learning and development Freeman, W. J. (1995). Societies of brains. Hillsdale, NJ: Erlbaum.
that stem from the physical body and from the envi- Garcia, J., and Koelling, R. A. (1966). Relation of cue to conse-
quence in avoidance learning. Psychonomic Science 4, 123124.
ronment. At the extreme, theorists advocating a dy-
Goldsmith, T. E., and Johnson, P. J. (1990). A structural assess-
namic systems approach claim that grounding cogni- ment of classroom learning. In R. Schvaneveldt, ed., Pathfind-
tion in the interaction of the body and the world er associative networks: Studies in knowledge organization. Nor-
obviates the need to propose mental representations wood, NJ: Ablex.
that mediate perception and action (Edelman, 1992; Johnson, M. (1987). The body in the mind: The bodily basis of meaning,
imagination, and reason. Chicago: University of Chicago Press.
Freeman, 1995; Johnson, 1987; Thelen and Smith,
Loftus, E. F., and Ketcham, K. (1991). Witness for the defense. New
1994; van Gelder, 1997). The grounding of concepts York: St. Martins.
in perception and action helps explain how concepts Meyer, D. E., and Schvaneveldt, R. W. (1976). Meaning, memory
are learned (Bickard, 2000). Consequently, coding is structure, and mental processes. Science 192, 2733.

Minsky, M. (1975). A framework for representing knowledge. In P. impaired subjects cognitive function has slipped back
Winston, ed., The psychology of computer vision. New York: onto the rising phase of the dose-response function.
Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Anrep. London:
In this case, treatments to facilitate cognition in im-
Oxford University Press. paired subjects act to restore optimal performance,
Quillian, M. R. (1969). The teachable language comprehender. but the same treatments in normal subjects drive
Communications of the ACM 12, 459476. them past the optimal level, resulting in impairment.
Schvaneveldt, R. W., Durso, F. T., and Dearholt, D. W. (1989). Net- The inverted-U hypothesis suggests that it is easier to
work structures in proximity data. In G. H. Bower, ed., The
psychology of learning and motivation: Advances in research and the-
treat memory deficits than to improve normal memo-
ory, Vol. 24. New York: Academic Press. ry. The results of many studies in animals and hu-
Smith, E. E., and Medin, D. L. (1981). Categories and concepts. Cam- mans support this idea.
bridge, MA: Harvard University Press.
Thelen, E., and Smith, L. B. (1994). A dynamic systems approach to Still, the possibility of memory enhancement in
the development of cognition and action. Cambridge, MA: MIT normal healthy people remains very attractive. A ca-
Press. sual search of the Internet leads to hundreds of sites
Tulving, E. (1962). Subjective organization in free recall of unre- offering memory-boosting aids. In some cases the
lated words. Psychological Review 69, 344354.
agents are described only as herbal mixtures, while in
van Gelder, T. (1997). Dynamics and cognition. In J. Haugland,
ed., Mind design II. Cambridge, MA: MIT Press. others a single compound, or one or more specific
components of the blend, is touted. The major con-
Roger W. Schvaneveldt
stituents of these products are discussed in detail later
in this entry. First, this entry reviews some drugs that
are known to be able to enhance cognition and mem-
ory but are not generally available or usually sold for
Virtually everyone has occasionally wished that he or
she could think faster, had a better memory, or was
CNS Stimulants
simply smarter. It is possible to improve these facets of
cognitive function through practice, but most people The common feature of these agents is that they
are looking for an easier solution. At the beginning stimulate the central nervous system. Picrotoxin,
of the twenty-first century, the answer is usually imag- bicuculline, and strychnine are drugs that block in-
ined to be in the form of a pill. hibitory systems in the brain, leading to a generalized
increase in excitability. Picrotoxin was one of the first
Cognition has many elements. Key aspects of cog-
agents recognized to be a memory enhancer. All three
nition include sensory perception, attention and con-
drugs have been shown in animal studies to improve
centration, immediate (or working) memory, and
memory at low doses, but higher doses cause seizures
long-term memory. More complex and integrated
and death. These drugs are too dangerous for experi-
facets of cognition include abstract reasoning and
mentation in humans.
planning. Thus, there are a number of ways for phar-
macological treatments to enhance cognition. Howev- Another stimulant drug, amphetamine, has reli-
er, for most people cognitive enhancement is synony- ably been shown to improve memory in both animals
mous with having a better long-term memory, so this and humans. Amphetamine causes the release of
topic is the primary focus of this entry. neurotransmitters that promote arousal, including
epinephrine and norepinephrine. The resulting in-
Memory itself is a complex phenomenon that is
crease in attention plays a major role in the memory-
still not completely understood by the medical and
enhancing properties of amphetamine. In addition,
scientific community. Much of the research into un-
amphetamine improves memory consolidation, the
derstanding memory mechanisms is motivated by the
process that leads to long-term memory storage. The
need to develop treatments for memory impairments
major problems with this drug are that tolerance to
caused by diseases such as Alzheimers disease. An im-
its beneficial effects develops quickly, along with side
portant question in the present context is whether
effects that include physical dependence and addic-
treatments that correct memory deficits can also en-
tion. These liabilities greatly outweigh amphet-
hance memory in normal healthy individuals. Virtual-
amines utility as a memory enhancer.
ly all drugs that improve memory show an invert-
ed-U dose-response relationship: Doses of a Caffeine is the worlds most widely used stimu-
compound up to a certain level enhance perfor- lant. It also increases alertness and attention, thereby
mance, but at still higher doses the enhancing effect enhancing cognitive performance. However, caf-
is lost. The argument has been made that cognitive feines effects on memory appear to be small. Vaso-
function in normal healthy subjects is at the crest of pressin is a pituitary hormone that plays an important
the inverted-U, while in aged, diseased, or otherwise role in the bodys regulation of water. In addition, va-

sopressin has been shown to enhance memory in both non of the inverted-U dosing effect, could explain the
humans and experimental animals. The precise inconsistent results of the few formal tests of these
mechanism is unknown, but since vasopressin affects agents that have been performed.
norepinephrine utilization in the brain researchers
believe that increased arousal plays a role. Five compounds seem to be most frequently sold
as cognition/memory enhancers: huperzine A, vin-
pocetine, Ginkgo biloba, ginseng, and pregnenolone.
Cholinergic Agents Huperzine A, an extract derived from a particular
type of club moss, is an acetylcholinesterase inhibitor
The neurotransmitter acetylcholine is known to
that is far more potent than physostigmine. The ace-
play an important role in learning and memory. Neu-
tylcholine plays an important role in the brain circuit-
rons that contain acetylcholine degenerate in patients
ry that encodes memories and is used as a neuro-
with Alzheimers disease, an illness that is character-
transmitter in the peripheral nervous system. Overac-
ized by profound cognitive impairments. In addition,
tivation of these peripheral cholinergic connections
drugs that block the action of acetylcholine in the
can have profound and potentially dangerous conse-
brain impair cognition in healthy humans and in ex-
quences. Because of this problem, huperzine A is not
perimental animals. Researchers have long hypothe-
a safe drug and should not be used without a physi-
sized that improving cholinergic system function
cians supervision.
would be a good way to treat the symptoms of Al-
zheimers disease, and this is the primary mechanism Vinpocetine is derived from the periwinkle plant,
of action of drugs that are marketed for this purpose. Cricoceras longiflorus. It is widely used as a memory en-
Cholinergic function can be enhanced either by hancer, and has been found to improve memory in
preventing the breakdown of the acetylcholine to pro- healthy people as well as those with impairments
long its action or by directly activating the receptors caused by aging or disease. It is not clear how vin-
for neurotransmitter. Physostigmine is an inhibitor of pocetine works, but there is evidence that vinpocetine
acetylcholinesterase, the enzyme that degrades ace- enhances blood flow in the brain, which, in turn, pro-
tylcholine. Physostigmine and related compounds vides the basis for observations of increased glucose
have been shown to improve memory in normal hu- utilization, a sign of generally increased cerebral ac-
mans and animals as well as in Alzheimers disease pa- tivity. In addition, studies have shown that vinpoce-
tients. Activators of either the muscarinic or nicotinic tine increases the production of norephinephrine,
receptor subtypes of cholinergic receptors also im- which mediates arousal, and acetylcholine. Vinpoce-
prove memory. Nicotine is well recognized as a cogni- tine is a very safe compound at therapeutic doses.
tion enhancer, although it is debated whether its ef-
Extracts from the leaves of the Ginkgo biloba tree
fect is mediated by increasing attention or through
have been used in traditional Chinese medicine for
another mechanism. Drugs that act at muscarinic re-
thousands of years. These extracts consist of many
ceptors have also been shown to be effective cognitive
compounds, the behavioral effects of which have yet
enhancers in animal studies. Unfortunately, several
to be individually characterized. Ginkgo extracts have
muscarinic receptor activators have failed in clinical
been widely used as memory enhancers, particularly
trials with Alzheimer patients because of side effects.
in the elderly, but studies have shown that ginkgo also
No compound in this class has yet been developed
improves memory in healthy young people. The pri-
that is safe enough for human use.
mary effect of ginkgo appears to be increased cere-
bral circulation.
Components of Smart Drugs The ginseng plant, Panax ginseng, also has a long
The vast number of concoctions sold over-the- history of use in traditional Chinese medicine. Ex-
counter to improve memory provides an excellent in- tracts from the roots or the entire plant contain over
dication of both the demand and the perceived need a dozen chemical compounds, collectively termed
for these products. Many of these compounds are ex- ginsenosides. Ginsenosides have been used to treat a
tracted from plants or animal tissues and consist of number of diseases ranging from diabetes to insom-
mixtures of ingredients. A major problem with such nia, and to enhance physical and mental perfor-
products is that claims concerning amounts and puri- mance. While the memory-enhancing activity of gin-
ty of the supposedly active elements have unusually seng is not as well documented as that of ginkgo,
not been verified by an independent agency. While studies have shown beneficial effects in both young
this lack of oversight could lead to safety problems and elderly subjects. Like ginkgo, ginseng also pro-
due to contaminants in the products, a more likely motes cerebral blood flow, but it also may enhance
concern is that dosages may not be accurate or consis- specific neurotransmitter systems. There is evidence
tent. Such unreliability, coupled with the phenome- that the combination of ginseng and ginkgo synergis-

tically enhance memory. There are no known major McGaugh, J. L., and Izquierdo, I. (2000). The contribution of
health risks associated with either agent. pharmacology to research on the mechanisms of memory for-
mation. Trends in Pharmacological Science 21 (6), 208210.
Pregnenolone is a neurosteroid, a hormone that Rees, K., Allen, D., and Lader, M. (1999). The influences of age
is synthesized in the brain from cholesterol. Studies and caffeine on psychomotor and cognitive function. Psycho-
have shown pregnenolone levels decline with aging pharmacology 145 (2), 181188.
Soetens, E., Casaer, S., DHooge, R., and Hueting, J. E. (1995). Ef-
and are correlated with impaired cognitive perfor- fect of amphetamine on long-term retention of verbal materi-
mance. In addition, an injection of pregnenolone di- al. Psychopharmacology 119 (2), 155162.
rectly into the brains of aged animals improves mem- Vallee, M., Mayo, W., and Le Moal, M. (2001). Role of pregneno-
ory. Pregnenolone has also been shown to enhance lone, dehydroepiandrosterone and their sulfate esters on
acetylcholine release in the brain. Taken together, learning and memory in cognitive aging. Brain Resolution Re-
view 37 (13), 301312.
these findings have been used to promote the value Warburton, D. M. (1995). Effects of caffeine on cognition and
of taking pregnenolone supplements. While pre- mood without caffeine abstinence. Psychopharmacology 119 (1),
gnenolone is safe, studies of the compound in hu- 6670.
mans have yet to provide consistent support for the Gregory M. Rose
idea that it is an effective cognitive enhancer.

Enhancing cognitive function remains both a
dream and a challenge. There are a number of agents Collective memory is a representation of the past that
that appear to at least improve memory to some de- is shared by members of a group, such as a generation
gree, but none of these compounds is so effective that or nation-state. The concept is usually traced to writ-
it can be distinguished from the others. What is clear ings of the French sociologist Maurice Halbwachs
is that some degree of cognitive enhancement is pos- (18871945), who argued that remembering is
sible even in young healthy people. Maximizing this shaped by participation in collective life and that dif-
effect will require systematic research to better under- ferent groups generate different accounts of the past
stand how pharmacological treatments affect the (Halbwachs, 1952).
basic neurobiological mechanisms that are responsi- Collective memory and related notions such as
ble for particular aspects of cognition. A more imme- public memory (Bodnar, 1992) have been examined
diate goal, and probably one that will be more quickly in academic disciplines including anthropology
achieved, is to develop treatments to improve the (Cole, 2001), history (Novick, 1999), and sociology
cognitive impairments in aged or diseased individu- (Schudson, 1992). Collective memory is also a part of
als. popular culture discussions about the Vietnam War
See also: DRUGS AND MEMORY; PHARMACOLOGICAL and the Holocaust. One of the hallmarks of collective
TREATMENT OF MEMORY DEFICITS memory is that it is tied to identity. Deeply held no-
tions about the past are often the source or pride or
Bibliography shame, and they can give rise to legal, and even
Born, J., Pietrowsky, R., and Fehm, H. L. (1998). Neuropsychologi- armed conflict. For example, the Serbs collective
cal effects of vasopressin in healthy humans. Progressive Brain memory of the Battle of Kosovo in 1389 has often
Research 119, 619643.
Furey, M. L., Pietrini, P., and Haxby, J. V. (2000). Cholinergic en-
been cited as a factor that made it possible to mobilize
hancement and increased selectivity of perceptual processing against Bosnian Muslims in the twentieth century.
during working memory. Science 290 (5,500), 2,3152,319.
Izquierdo, I., and Medina, J. H. (1991). GABAA receptor modula-
tion of memory: The role of endogenous benzodiazepines. Strong and Distributed Accounts of
Trends in Pharmacological Science 12 (7), 260265. Collective Memory
Kennedy, D. O., Scholey, A. B., and Wesnes, K. A. (2000). The
dose-dependent cognitive effects of acute administration of Collective memory is often understood in terms
Ginkgo biloba to healthy young volunteers. Psychopharma- of loose analogies with memory in the individual.
cology, 151 (4), 416423. Many discussions of Americas memory about Viet-
(2001). Dose dependent changes in cognitive performance nam, for example, seem to presuppose that America
and mood following acute administration of Ginseng to is some sort of a large organism that has intentions,
healthy young volunteers. Nutritional Neuroscience 4 (5), 295
310. desires, memories, and beliefs just as individuals do,
Kidd, P. M. (1999). A review of nutrients and botanicals in the inte- something reflected in assertions such as, Our mem-
grative management of cognitive dysfunction. Alternative Med- ory of Vietnam makes us unwilling to accept combat
icine Review 4 (3), 144161. deaths.
Mattay, V. S. et al. (2000). Effects of dextroamphetamine on cogni-
tive performance and cortical activation. Neuroimage 12 (3), Assumptions about this issue are often not well
268275. grounded and as a result have been the object of criti-

cism. For example, one of the fathers of modern sphere (Bodnar, 1992; Wertsch, 2002). Reflecting its
memory studies in psychology, Frederic Bartlett ties to identity, it is often assessed in terms of its ability
(18861969), was critical of the more or less absolute to provide a usable past, even when this representa-
likeness [that] has been drawn between social groups tion comes at the expense of accuracy. In contrast,
and the human individual (Bartlett, p. 293), and he while recognizing a connection to identity studies of
warned that collectives do not have some sort of mem- individual memory tend to focus on a criterion of ac-
ory in their own right. Bartlett did believe, however, curacy. This is not to say that psychologists believe
that memory of individuals is fundamentally influ- memory is fundamentally accurate. Indeed, many
enced by the social context in which they function. In- studies point to the myriad ways humans can generate
deed, a central point of his argument is that social incorrect accounts of what actually occurred. The
organization gives a persistent framework into which point remains, however, that psychological studies
all detailed recall must fit, and it very powerfully in- formulate memory largely in terms of its degree of ac-
fluences both the manner and the matter of recall (p. curacy.
296). In short, Bartlett accepted the notion of mem- A question that is sometimes asked about collec-
ory in the group, and not memory of the group (p. tive memory is whether it is anything more than a set
294). of individual memories and hence whether it could be
Claims about memory of the group constitute reduced to a personal level. Any attempt to answer
a strong version of collective memory (Wertsch, this question must say something about the social pro-
2002), and, when made explicit, they have usually cesses involved in collective memory without falling
been rejected. An alternative that recognizes memory into a strong version. This is often done by focusing
in the group without slipping into questionable as- on how narratives and other textual resources for
sumptions about memory of the group is a distribut- memory are produced, discussed, and understood by
ed version. From this perspective, memory is viewed members of a collective.
as being distributed socially in small group interac- For example, modern states devote major atten-
tion; as well as instrumentally (Wertsch, 2002). In tion to these issues. They produce official accounts of
the case of social distribution, for example, Mary Sue the past through textbooks and other materials used
Weldon (2001) has examined the collaborative re- in schools, national commemorations, the media, and
membering that occurs when groups of individuals other forms of popular culture. In carrying out this
work together to recall information or events from project, states seek not only to promulgate their own
the past. account but also to control access to alternative ac-
In the case of instrumental distribution, memory counts. As depicted by George Orwell in his futuristic
is viewed as involving an active agent and one or more novel 1984 (1949), such control can be almost total,
cultural tools such as calendars or other written re- but in fact every modern state exerts some control
cords. An important transformation of memory in over the account of the past presented to its citizens.
human cognitive evolution occurred with the emer- This focus on the textual mediation of collective
gence of external symbolic storage (Donald, 1991). memory raises the question of whether collective
This does not mean that such memory somehow re- memory is really memory at all. What does it mean to
sides in texts or records, but it does mean that with the say, for example, that a generation of American teen-
rise of new forms of external symbolic storage such as agers remembers World War II when they were born
written texts or the Internet the possibilities for re- four decades after the event? This might involve
membering undergo fundamental change. memory for textual resources, or perhaps semantic
Such change has both psychological and social di- memory, but is it really the kind of episodic memory
mensions. By becoming skilled at using a certain set often assumed?
of cultural tools, new mental habits and schemata Such questions provide a reminder of the prob-
(Bartlett, 1995) emerge that shape remembering for lems that arise when loose analogies between individ-
members of a collective. New forms of collective life ual and collective remembering are employed. In
also derive from the appearance of novel cultural order to address them, one must again turn to the in-
tools. For example, Benedict Anderson (1991) has strumental distribution of collective memory, and in
tied the development of modern nations and other the end it may be more appropriate to discuss this in
imagined communities to the rise of print media. terms of knowledge about a set of shared textual re-
sources rather than memory per se. What is striking
about collective memory, however, is the ways in
Collective versus Individual Memory which it goes beyond what would normally be de-
Collective memory is widely understood as inher- scribed as knowledge or semantic memory. This is so
ently involving conflict and negotiation in the social first because of the social conflict and control involved

in producing the textual resources, and second be- A final property that characterizes collective re-
cause much more than neutral knowledge is involved. membering is that it tends to be heavily shaped by
When discussing highly charged historical events, it schemata (Bartlett, 1995), implicit theories
is very easy to slip into heated discussions about what (Ross, 1989), or other simplifying and organizing
really happened and deny others accounts in such frameworks. Such frameworks also shape individual
a way that narrative texts become emotionally power- memory and history, but in the case of collective
ful tools of collective identity and memory. memory they take on a particularly important role
One of the ways that studies of individual and col- due to the processes of conflict and negotiation in-
lective memory have overlapped concerns genera- volved. What is shared in collective memory is often
tional differences. Martin A. Conway (1997) and oth- little more than a schematic narrative template
ers have examined the reminiscence bump that (Wertsch, 2002) in which detailed information, espe-
exists for members of a generation for events that oc- cially contradictory information, is distorted, simpli-
curred when they were between fifteen and twenty- fied, and ignored; this schema stands in contrast to
five years of age. Research suggests that such events what are at least the aspirations of analytic history. It
provide the foundation for a generations lifelong col- is for this reason that collective memory often appears
lective memory. Providing another hint about the to be impervious to new information that might chal-
close tie between identity and collective memory, lenge it.
Conway argues that the reminiscence bump and its Bibliography
power to shape a generation are inherently tied to Anderson, B. (1991). Imagined communities: Reflections on the origin
processes of identity development in young adult- and spread of nationalism. London: Verso.
hood. Bartlett, F. C. (1932; reprint 1995). Remembering: A study in experi-
mental and social psychology. Cambridge, UK: Cambridge Uni-
versity Press.
Collective Remembering versus History Bodnar, J. (1992). Remaking America: Public memory, commemoration,
and patriotism in the twentieth century. Princeton, NJ: Princeton
Notions of collective memory and history often University Press.
overlap. In both cases, the events involved are likely Chang, I. (1997). The rape of Nanking: The forgotten holocaust of World
to have occurred before the lifetime of the individuals War II. New York: Basic Books.
Cole, J. (2001). Forget colonialism? Sacrifice and the art of memory in
or group doing the remembering, and in both cases
Madagascar. Berkeley: University of California Press.
there is a tendency to assert that the account being Conway, M. A. (1997). The inventory of experience: Memory and
presented is true. Furthermore, both rely on narra- identity. In J. W. Pennebaker, D. Paez, and B. Rim, eds., Col-
tive as a cultural tool. Despite such shared properties, lective memory of political events: Social psychological perspectives.
it is possible, indeed essential, to distinguish between Mahwah, NJ: Erlbaum.
Donald, M. (1991). Origins of the modern mind: Three stages in the evo-
lution of culture and cognition. Cambridge, MA: Harvard Uni-
Among the properties of collective memory that versity Press.
tend to distinguish it from history are the following: Halbwachs, M. (1952; reprint 1992). On collective memory, ed. and
trans. Lewis A. Coser. Chicago: University of Chicago Press.
1. it reflects a single, subjective, committed perspec- Novick, P. (1999). The Holocaust in American life. Boston: Houghton
tive of a collective, whereas analytic history strives Mifflin.
Ross, M. (1989). Relation of implicit theories to the construction
to be objective and distance itself from any partic- of personal histories. Psychological Review 96 (2), 341357.
ular perspective; Schudson, M. (1992). Watergate in American memory: How we remem-
2. collective memory leaves little room for doubt or ber, forget, and reconstruct the past. New York: Basic Books.
Weldon, M. S. (2001). Remembering as a social process. In D. L.
ambiguity about events and the motivations of ac-
Medin, ed., The psychology of learning and motivation. San
tors (Novick, 1999), whereas analytic history Diego, CA: Academic Press.
strives to take into account multiple, complex fac- Wertsch, J. V. (2002). Voices of collective remembering. New York:
tors; Cambridge University Press.

3. collective memory presupposes an unchanging James V. Wertsch

essence of a group across time, whereas analytic
history focuses on the transformations that collec-
tives undergo.
These three properties of collective memory charac-
terize, for example, the highly charged dispute be- Comparative cognition is the comparison of how ani-
tween China and Japan over whether events in 1937 mals (including humans) process and interact with
constitute the rape of Nanking (Chang, 1997) or the world. What sets animal cognition apart from the
the incident in Nanking (as mentioned in some behavioristic tradition of stimulus-response (S-R)
Japanese textbooks). learning is the recognition that (mental) processes in-

Figure 1

An information-processing model of temporal processing representative of scalar timing theory.

tervene between the stimulus and response. Some an- creases. The peak indicates maximum expectancy of
imal-cognitive procedures may be complex (e.g., reward. Interval timing typically displays the scalar
feeling of knowing experiments by Hampton, propertyvariability (spread) increases proportion-
2001). But even simple Pavlovian conditioning pro- ally with time. The interval-timing theory receiving
cedures can deal with cognitive processes: for exam- the most attention is shown in Figure 1. It includes a
ple, taste-aversion conditioning to the image of an clock (pacemaker), a switch (gate) to an accumulator
expected type of food (e.g., Holland, 1990), or (integrator), reference memory for the expected
higher-order and backward fear conditioning to a time, and a comparator for the reference time versus
neutral stimulus never paired with shock (Cole, Bar- accumulated time to see if times up. Converging
net, and Miller, 1995). Techniques developed in the evidence has been accumulating for each of these hy-
late twentieth century to study cognitive processes pothetical processes. One example is that drug ma-
have produced exciting revelations about how ani- nipulations of the clock produce abrupt changes in
mals perceive, learn, remember, navigate, communi- peak time (see Figure 2) with gradual compensation
cate, time, and count. This entry highlights some of by storage of new accumulator values in reference
these developments. memory, whereas drug manipulations of the memory
produce gradual changes (see Figure 3) due to dis-
torted intervals accumulating in reference memory
Interval Timing (Meck, 1996).
A tremendous amount of information has accu-
mulated since the 1980s about the interval-timing
mechanisms of rats, pigeons, and primates. In a pop- Numerosity: A Scale of Numbers
ular peak procedure, a stimulus cue is presented Learning a number scale is fundamental to math-
and after a fixed elapsed time the next response is ematics. Evidence indicates that rhesus monkeys can
rewarded. On some trials, reward is omitted and learn a number scale (Brannon and Terrace, 1998,
responding increases to a peak rate and then de- 2000). Monkeys touched randomly placed areas on a

Figure 2

Mean peak times from four different groups of rats (n=10/group) initially trained on either a 40-s (squares) or 20-s (circles) PI-timing
procedure and tested under the influence of dopaminergic drugs that affect clock speed (left panel) or after the removal of the drugs
(right panel). The closed symbols represent rats treated with methamphetamine (1.5 mg/kg i.p.), and the open symbols represent rats
treated with haloperidol (0.12 mg/kg i.p.).

computer screen in order: first the area containing Landmark Navigation and Cognitive Maps
one object, then two, then three, and finally four ob- Clarks nutcrackers apparently use directional
jects. Objects varied in size, shape, or color and com- bearings from several landmarks to fix the location
binations of these dimensions within and across trials. of seed caches (Kamil and Jones, 2000; Kamil and
The most remarkable result was that this training Cheng, 2001). Nutcrackers were more accurate using
spontaneously generalized to larger novel numbers of multiple bearings than any single bearing with
five to nine objects beyond the range of training num- a distance measure along the bearing line. This
bers. remarkable finding came from manipulating the

Figure 3

Mean peak times from four different groups of rats (n=10/group) initially trained on either a 40-s (squares) or 20-s (circles) PI-timing
procedure and tested under the influence of cholinergic drugs that affect memory-storage speed (left panel) or after the removal of the
drugs (right panel). The closed symbols represent rats treated with physostigmine (0.01 mg/kg i.p.) and the open symbols represent
rats treated with atropine (0.05 mg/kg i.p.).

goal site in relationship to geometrical properties than configural-landmark navigation (Shettleworth,

of fixed landmarks rather than manipulating the 1998).
landmarks themselves. Landmarks seem to function
as a configural whole, not as independent elements.
Young children, but not older children or adults, use Abstract Concepts
similar geometric navigation strategies (Hermer- Researchers have long debated which species are
Vazquez, Moffet, Munkholm, 2001). The issue of capable of abstract concept learning. Abstract con-
cognitive maps, however, is complex and is more cepts are rules (same versus different, identity match-

ing) that transcend the training stimuli. Novel stimuli ented with pairs of them (behind a window). They
are used to test for abstract concept learning. Pigeons were required to pull one of two ropes to indicate
(like humans, apes, dolphins, and monkeys) are capa- whether they were of the same or different categories
ble of learning abstract concepts. One study showed (food versus nonfood). Novel item transfer was better
matching-to-sample (MTS) concept learning by pi- than 80 percent correct. Rhesus monkeys have also
geons (Wright, 1997). In MTS, a sample stimulus is showed category matching (Neiworth and Wright,
followed by two choice stimuli, one of which is correct 1994).
and matches the sample. Pigeons required to respond
to (peck) the sample twenty times before getting the
choice stimuli learned the concept fully (as opposed Relations Between Relations: Analogical
to partially). Other pigeons required to respond fewer Reasoning
times (e.g., once) did not learn the concept at all, and If abstract concept learning is higher order, then
instead learned the configural pattern of each display analogical reasoning is higher-higher order. The Mil-
(sample plus choice stimuli). lers Analogies Test is a test of relations between rela-
tions. A recent MTS study showed that baboons could
Pigeons are also capable of learning a same/
learn a relation between relations (Fagot, Wasserman,
different (S/D) abstract concept. In S/D, subjects iden-
and Young, 2001). Baboons saw a brief sixteen-icon
tify stimuli as same or different. Pigeons most rapidly
sample, which was then replaced by two sixteen-icon
learn S/D (and the S/D concept partially) with large
choice displays. Neither choice display had any icons
384 element arrays and contrasting target regions on
in common with the sample. The baboons successfully
different trials (Cook, Katz, and Cavoto, 1997), or
learned and transferred this performance to novel
with sixteen-element arrays that are composed of all
stimuli, which could not be based on stimulus identity
different or all same elements (Young, Wasserman,
or functional categories. It is difficult to escape the
and Garner, 1997). But when the number of elements
conclusion that this behavior is based on a relation be-
in these tasks is reduced to two (minimum number),
tween relations. A similar conclusion came from a
performance falls to chance. Nevertheless, pigeons
study on rhesus monkey music perception (Wright et
are capable of learning a S/D task when trained with
al., 2000). Six-note tonal (childhood songs, tonal al-
two stimuli from the beginning and attain full S/D
gorithm) musical passages, separated by one or two
concept learning when the training set is expanded
octaves were judged same better than 80 percent of
to 128 to 512 pictures. Rhesus monkeys attained full
the time. Tonal passages form a relation (tune), and
S/D concept learning with somewhat smaller set sizes
thus this octave generalization (no common frequen-
than pigeons, and chimpanzees remarkably learned
cies) is a relation between relations.
an S/D concept with only two training stimuli (Oden,
Thompson, and Premack, 1988). Great apes are particularly remarkable in their
ability to form relations between relations. Chimpan-
Equally remarkable is the S/D concept learning zees with no relations-between-relations training
by a parrot, Alex, with stimuli that varied along three spontaneously performed correct analogical
dimensions (Pepperberg, 1987). Alex responded with choices without reward (Thompson, Oden, and
the English words color, shape, or mah-mah Boysen, 1997). If anything could be more remark-
(matter). When Alex was presented with a red wooden able, Sarah, a highly trained chimpanzee arranged
triangle and a green rawhide triangle, for example, four out of five scattered geometric forms in the
and asked, Whats same? he typically said, Shape. unique analogical relationship (and set aside the dis-
When presented with a red wooden square and a blue tractor) without explicit training (Thompson and
wooden square and asked, Whats different? he typ- Oden, 2000).
ically said, Color. Alexs novel object transfer per-
formance was 85 percent correct. What makes this ex-
periment so remarkable was that Alex had no way of Episodic Memory
telling which dimension (color, shape, or matter) Episodic memory is memory for a particular
would be questioned and whether he would be asked, event and is related to the issue of consciousness in
Whats same? or Whats different? animals. Recent experiments show that scrub jays
have episodic memory (Clayton and Dickinson, 1998;
Clayton and Dickinson, 1999; Emery and Clayton,
Abstract Concepts Based on Categories 2001). In one experiment, jays cached wax worms and
Research developments have shown that mon- peanuts in separate and distinctive halves of two dis-
keys could learn an S/D concept with food versus non- tinctive sand-filled ice cube trays (two-by-seven ar-
food categories (Bovet and Vauclair, 2001). Baboons rays). After four hours, the jays recovered the more
were given experience with the items and then pres- desirable wax worms first from one tray. After 124

hours, they recovered the peanuts first from the other ditions would not show this ability. The study of quan-
tray. In pretraining, they had learned that the wax titative differences encounters the same difficulty.
worms deteriorated after 124 hours. Reversal of their Some argue that comparisons should be among close-
earlier preference for wax worms means that they re- ly related species to study cognitive evolution (Cam-
membered what (peanuts versus wax worms), when bell and Hodos, 1991) or environmentally specialized
(four versus 124 hours), and where (which tray side) cognitive behavior (Sherry and Schacter, 1987). But
the foods were stored. Cache recovery by Clarks nut- even testing closely related species (e.g., Clarks nut-
crackers in the forest must also be episodic memory crackers, Mexican jays, pinyon jays, and scrub jays) in
because they use different cache sites each year and the same task does not ensure that quantitative differ-
overcome potential proactive interference (i.e., con- ences represent cognitive differences, as opposed to
fusions) from previous years sites. Thus, they too the task fitting the predispositions of some of the spe-
know what, when, and where. cies better. Nevertheless, testing the species in a vari-
ety of tasks and varying critical parameters can con-
verge on persuasive evidence for quantitative
Feeling of Knowing: Metacognition
differences (Olson, Kamil, Balda, and Nims, 1995).
Exciting research developments with monkeys in-
dicate that they know when they remember (Hamp- A more overarching goal would seem to under-
ton, 2001). This ability in humans indicates conscious stand the mechanisms and processes involved; in
cognition and is closely aligned with the issue of con- short, how cognition works. Any thorough under-
sciousness and recognition of self. In a MTS experi- standing will need evidence from at least three do-
ment, rhesus monkeys saw a sample (clip-art image), mains: conditions (i.e., natural context) best suited to
touched the sample, had a retention delay, and then express this cognitive ability; laboratory tests that ma-
decided whether or not to take a memory test. If they nipulate critical parameters over a substantial range;
chose to take the memory test, four choices were pres- and neurobiology and neural circuitry that mediates
ented on the four corners of the monitor and reward this behavior. One can begin in any domain, but will
for the correct choice was highly desired peanuts. If need regular input from the other domainsa sort of
they declined to take the test, they (automatically) re- cumulative upward spiral and a revisit of issues as evi-
ceived a less preferred pellet. The monkeys were dence accumulates. Ultimately, quantitative differ-
more accurate when they chose to take the test than ences (e.g., interval timing [Church, Meck, and Gib-
when they were required to take the test, and this dif- bon, 1994]) and even qualitative differences (e.g.,
ference increased with delay. An important aspect of serial-order learning [Terrace, Chen, and Newman,
this study was requiring the monkeys to monitor their 1995]; memory rehearsal [Cook, Wright, and Sands,
memory before the test was presented. It is difficult 1991]) should be shown to be individual points of
to escape the conclusion that these monkeys know comparison along much larger functional relation-
when they remember. Pigeons do not show this ships revealed by parametric studies and quantitative
choice-test advantage (Inman and Shettleworth, models.
1999), and thus may lack metacognition. The Cognitive Issue
Issues associated with the cognition part of com-
Directions for Future Research parative cognition mainly revolve around the
Animals have cognitive abilities that were previ- strength of evidence for private events (cognitive pro-
ously thought to be uniquely human. Other qualita- cesses) between the stimulus and response. There will
tive differences may disappear as procedures better certainly be skeptics of the feeling-of-knowing experi-
fit the predispositions of species being tested (Clayton ments with animals (Hampton, 2001), but this is only
and Dickinson, 1999). But there are issues other than the tip of the iceberg. It gets only better (or worse de-
finding human cognitive abilities in animals and pending on ones persuasion) with experiments on
these issues continue to revolve around what is meant beliefs, intentions, wants, desires, altruism, concept of
by comparative and cognition. self, and knowledge about anothers beliefs and inten-
tionsthe so-called theory of mind experiments
The Comparative Issue (Povinelli, 2001; Tomasello and Call, 1997). These
Criticisms of comparative cognition have been investigations have stimulated a great deal of enthusi-
that they are not comparative (Shettleworth, 1993). asm among scientists. Continued growth and accep-
However, researchers have seen species comparisons tance will likely depend on parametric studies (i.e.,
in interval timing, landmark navigation, abstract con- manipulating critical variables over a substantial
cept learning, analogical reasoning, and metacogni- range) and converging evidence from several experi-
tion. A problem with proving qualitative differences ments like the blue jay and wax-worm studies. Fur-
is that one can never be sure that some other test con- thermore, researchers will need to show that their

findings cannot be accounted for by simpler associa- Hermer-Vazquez, L. Moffet, A., and Munkholm, P. (2001). Lan-
tive learning principles involving observable stimuli guage, space, and the development of cognitive flexibility in
humans: The case of two spatial memory tasks. Cognition 79,
and responses, like behavior-based theories timing 263299.
(Killeen and Fetterman, 1988; Machado, 1997) as op- Holland, P. C. (1990). Event representation in Pavlovian condi-
posed to scalar-timing theory (see Figure 1). A wide tioning: Image and action. Cognition 37, 105131.
variety of learning, cognitive, and memory phenome- Inman, A., and Shettleworth, S. J. (1999). Detecting metamemory
na can be accounted for by careful parametric experi- in nonverbal subjects: A test with pigeons. Journal of Experi-
mental Psychology: Animal Behavior Processes 25, 389395.
mentation coupled with powerful mathematical mod-
Kamil, A. C., and Cheng, K. (2001). Way-finding and landmarks:
els to reveal underlying processes (Gallistel and The multiple-bearings hypothesis. Journal of Experimental Biol-
Gibbon, 2000; White and Wixted, 1999). Any time we ogy 204, 103113.
are tempted to get too heady over some new animal- Kamil, A. C., and Jones, J. E. (2000). Geometric rule learning by
cognitive phenomenon, we might reflect on the exu- Clarks nutcrackers (Nucifraga columbiana). Journal of Experi-
mental Psychology: Animal Behavior Processes 26, 439453.
berance originally generated by ape-language
Katz, J. S., Wright, A. A., and Bachevalier, J. (2002). Same/different
studies, and recall the instructive (if not tongue-in- concept learning in Rhesus monkeys. Journal of Experimental
cheek) S-R conditioning demonstrations of the pi- Psychology: Animal Behavior Processes 28.
geons (cognitive) concepts of insight, symbolic com- Killeen, P. R., and Fetterman, J. G. A behavioral theory of timing.
munication, and self conducted by Epstein, Skinner, Psychological Review 95, 274295.
and their collaborators (1981). Machado, A. (1997). Learning the temporal dynamics of behavior.
Psychological Review 104, 241265.
Meck, W. H. (1996). Neuropharmacology of timing and time per-
Bibliography ception. Cognitive Brain Research 3, 227242.
Neiworth, J. J., and Wright, A. A. (1994). Monkeys (Macaca mulat-
Bovet, D., and Vauclair, J. (2001). Judgment of conceptual identity
in monkeys. Psychonomic Bulletin & Review 8, 470475. ta) learn category matching in a nonidentical same-different
Brannon, E. M., and Terrace, H. S. (1998). Ordering of the task. Journal of Experimental Psychology: Animal Behavior Process-
numerosities 1 to 9 by monkeys. Science 282, 746749. es 20, 429435.
(2000). Representation of the numerosities 19 by rhesus Oden, D. L., Thompson, R. K., and Premack, D. (1988). Spontane-
macaques (Macaca mulatta). Journal of Experimental Psychology: ous transfer of matching by infant chimpanzees (Pan troglo-
Animal Behavior Processes 26, 3149. dytes). Journal of Experimental Psychology: Animal Behavior Pro-
Campbell, C. B. G., and Hodos, W. (1991). The scala naturae revis- cesses 14, 140145.
ited: Evolutionary scales and anagenesis in comparative psy- Olson, D. J., Kamil, A. C., Balda, R. P., and Nims, P. J. (1995). Per-
chology. Journal of Comparative Psychology 105, 211221. formance of four seed-caching corvid species in operant tests
Church, R. M., Meck, W. H., and Gibbon, J. (1994). Application of of nonspatial and spatial memory. Journal of Comparative Psy-
scalar timing theory to individual trials. Journal of Experimental chology 109, 173181.
Psychology: Animal Behavior Processes 20, 135155. Pepperberg, I. M. (1987). Acquisition of the same/different concept
Clayton, N. S., and Dickinson, A. (1998). Episodic-like memory by an African Grey parrot (Psittacus erithacus): Learning with
during cache recovery by scrub jays. Nature 395, 272274. respect to categories of color, shape, and material. Animal
(1999). Memory for the content of caches by scrub jays Learning and Behavior 15, 423432.
(Aphelocoma coerulescens). Journal of Experimental Psychology: Povinelli, D. J. (2001). On the possibilities of detecting intentions
Animal Behavior Processes 25, 8291. prior to understanding them. In B. F. Malle, L. J. Moses, and
Cole, R. P., Barnet, R. C., and Miller, R. R. (1995). Temporal en- D. A. Baldwin, eds., Intentions and intentionality: Foundations of
coding in trace conditioning. Animal Learning and Behavior 23, social cognition. Cambridge, MA: MIT Press.
144153. Sherry, D. F., and Schacter, D. L. (1987). The evolution of multiple
Cook, R. G., Katz, J. S., and Cavoto, B. R. (1997). Pigeon same- memory systems. Psychological Review 94, 439454.
different concept learning with multiple stimulus classes. Jour- Shettleworth, S. J. (1993). Where is the comparison in comparative
nal of Experimental Psychology: Animal Behavior Processes 23, cognition? Alternative research programs. Psychological Science
417433. 4, 179184.
Cook, R. G., Wright, A. A., and Sands, S. F. (1991). Interstimulus (1998). Cognition, evolution, and behavior. New York: Oxford
interval and viewing time effects in monkey list memory. Ani- University.
mal Learning and Behavior 19, 153163. Terrace, H. S., Chen, S., and Newman, A. B. (1995). Serial learning
Emery, N. J., and Clayton, N. S. (2001). It takes a thief to know a with a wild card by pigeons (Columba livia): Effect of list
thief: Effects of social context on prospective caching strate- length. Journal of Comparative Psychology 109, 1,1621,172.
gies in scrub jays. Nature 414, 443446. Thompson, R. K. R., and Oden, D. L. (2000). Categorical percep-
Epstein, R., Lanza, R. P., and Skinner, B. F. (1981). Self- tion and conceptual judgments by nonhuman primates: The
awareness in the pigeon. Science 212, 695696. paleological monkey and the analogical ape. Cognitive Science
Fagot, J., Wasserman, E. A., and Young, M. E. (2001). Discriminat- 24, 363396.
ing the relation between relations: The role of entropy in ab- Thompson, R. K. R., Oden, D. L., and Boysen, S. T. (1997). Lan-
stract conceptualization by baboons (Papio papio) and hu- guage-nave chimpanzees (Pan troglodytes) judge relations
mans (Homo sapiens). Journal of Experimental Psychology: between relations in a conceptual matching-to-sample task.
Animal Behavior Processes 27, 316328. Journal of Experimental Psychology: Animal Behavior Processes 23,
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition- 3143.
ing. Psychological Review 107, 289344. Tomasello, M., and Call, J. (1997). Primate cognition. New York: Ox-
Hampton, R. R. (2001). Rhesus monkeys know when they remem- ford University Press.
ber. Proceeding of the National Academy of Sciences of the United White, K. G., and Wixted, J. T. (1999). Psychophysics of remember-
States of America 98, 5,3595,362. ing. Journal of the Experimental Analysis of Behavior 71, 91113.

Wright, A. A. (1997). Concept learning and learning strategies. Psy- Medin, 1981) have led to a shift in attention from the
chological Science 8, 119123. classical view to the probabilistic view.
Wright, A. A., Rivera, J. J., Hulse, S. H., Shyan, M., and Neiworth,
J. J[GU2]. (2000). Music perception and octave generalization The probabilistic view argues that most concepts
in rhesus monkeys. Journal of Experimental Psychology: General are organized around properties that are only charac-
129, 291307. teristic or typical of a category (rather than defining).
Young, M. E., Wasserman, E. A., and Garner, K. L. (1997). Effects
One specific account of the probabilistic view assumes
of number of items on the pigeons discrimination of same
from different visual displays. Journal of Experimental Psycholo- that a concept is a summary representation or proto-
gy: Animal Behavior Processes 23, 491501. type that indicates what is, on the average, true of a
category (e.g., the bird prototype would include fea-
Anthony A. Wright
tures such as flies and sings because the properties are
true of most though not all birds). The prototype view
readily handles goodness-of-example effects; the
more similar an item is to the prototype, the more
CONCEPTS AND CATEGORIES, typical it will be of the category. For example, robins
LEARNING OF would be more similar to the bird prototype than are
Concepts are a fundamental aspect of intelligent be- ostriches because they have more features that are
havior. Traditionally, a concept has been viewed as a generally true of birds. Thus, they would be better ex-
mental representation that picks out a group of equiv- amples of birds.
alent items or a category. For example, every person However, the prototype view also has problems.
has a concept of dog and can use that concept to pick Prototype representations alone are not rich enough
out a category of things that one would call dogs. Some to capture peoples knowledge about a category. For
of the most fundamental questions about the mind in- example, people are sensitive to the number of in-
clude the following: What do human concepts consist stances in a category (e.g., there are many more hous-
of (i.e, what is their structure)? How are they are ac- es than igloos), the variability of features (e.g., the size
quired? Why do humans have concepts (i.e, What of quarters varies less than the size of pizzas), and the
functions do they have)? correlations between features (e.g., wooden spoons
tend to be big whereas metal spoons tend to be small
[Medin, 1989]). To overcome such difficulties, re-
What Do Human Concepts Consist Of?
searchers have suggested that instead of (or in addi-
An early, popular view of concepts was the classi- tion to) a prototype people may simply store memory
cal view. For a variety of reasons this approach was instances or examples of the category and reason with
unsatisfactory and gave way to the probabilistic view. them (Brooks, 1978). Thus, according to the exem-
These views are described and compared below. plar view, for instance, the concept of a chair would
Classical versus Probabilistic View include memory traces of particular chairs and their
associated features.
The classical view argues that concepts are struc-
tured around defining features (Bruner, Goodnow, The Theory View
and Austin, 1956). Defining features are features that Another approach to concepts is the theory view.
are singly necessary and jointly sufficient to define the As described below, this approach developed as a re-
concept. For example, the concept bachelor has the action to certain limitations of the classical and proba-
defining features human, unmarried, and male. bilistic approaches. In particular, the classical and
However, the classical view appears to have a probabilistic views failed to take into account peoples
number of serious problems. First, if concepts have background knowledge or theories of the world.
defining features, then one ought to be able to specify The shift from the classical view to the probabilis-
what they are. But many common concepts, such as tic view was motivated by a detailed analysis of natural
game or chair, seem to have no defining features. In- object categories. Associated with this analysis is the
stead, instances of these concepts have characteristic view that concepts or mental representations of cate-
features that are neither necessary nor sufficient for gories closely mirror the structure afforded by prop-
category membership (e.g., has four legs and has a back erties of category members. It seems almost a tautolo-
for chairs). Second, not all instances of a concept are gy that if the structure of examples does not have
equally good examples of that concept. For example, defining features then the corresponding mental rep-
people judge robins to be better examples of the con- resentations cannot conform to the classical view.
cept bird than ostriches. If both robins and ostriches Similarly, a probabilistic category structure suggests
have the defining features of birds, then why should a probabilistic concept representation. In brief, re-
robins be considered better examples of birds than searchers assume that mental representations are de-
ostriches? These and other problems (Smith and termined by the structure of examples in the world.

However, the classical and probabilistic views bers (Posner and Keele, 1968), by attending to fea-
tend to ignore the role of the learner in their accounts tures commonly shared by members and discarding
of concepts. According to the theory view, learners features varying among members (Elio and Ander-
also impose structure on their concepts. That is, con- son, 1981), or by noting the most common value on
cepts are based on a learners general knowledge and each dimension. The basic idea behind these models
theories of the world together with information pro- can be traced to Galtons composite photograph
vided by the environment (Carey, 1985; Murphy and theory (Galton, 1879). Galton superimposed several
Medin, 1985; Rips, 1989). For example, Susan Carey faces to make a composite photograph in which com-
showed that childrens biological theories influence mon properties were accentuated and variant proper-
their patterns of inductions at a very early age. To il- ties were attenuated. Such a process is assumed in
lustrate, a mechanical monkey is rated by both chil- prototype theories. On the other hand, exemplar
dren and adults to be more similar to a human being models assume category instances are stored but gen-
than is a worm, yet even young children infer that erally have not specified detailed learning mecha-
worms rather than toy monkeys have a spleen after nisms (see Kruschke, 1992, for an exception). These
being told that people have a spleen, a round and green views assume that the learner begins with features of
[thing] . . . in the persons body. In this example, the fea- the entities and then learns which features are impor-
ture has a spleen, which is more consistent with a tant for the concept. However, research conducted in
childs background knowledge or theory about ani- the 1990s suggests that an important part of concept
mate things than inanimate things. learning is learning to identify the features them-
Theories themselves may be anchored by how selves (Schyns, Goldstone, Thibaut, 1998).
well their predictions receive support from the world. The theory-based view of concepts takes a differ-
Gregory Murphy and Douglas Medin (1985) suggest ent perspective on concept formation. Several re-
that the relation between concept and examples is searchers have proposed that humans may be born
like that between theory and data. Thus, concepts with a naive physics and a naive biology or psychology
would not necessarily consist of features that are also (Carey, 1985; Keil, 1989; Spelke, 1990) that act as ini-
in examples; rather, the constituents of examples tial theories to organize conceptual knowledge. A
would only need to support more abstract constituents major implication of the theory-based view is that
of concepts (Wisniewski and Medin, 1994). For exam- concept learning involves integrating new examples
ple, one may infer that a man is drunk because one with prior knowledge. In particular, prior knowledge
sees him jump into a pool fully clothed. If one does may influence the identification of features and, in
so, it is probably not because the feature jumps into turn, information about examples may modify a per-
pools, clothed is listed with the concept drunk. Rather, sons prior knowledge (Wisniewski and Medin, 1994).
it is because part of ones concept of drunk involves a
theory of impaired judgment that serves to explain Taking the theory-based view, a group of re-
the mans behavior. searchers in artificial intelligence (an area in comput-
er science the goal of which is to develop computers
to do intelligent things) have developed models of
How Are Concepts Acquired? concept formation called explanation-based learning
Young children probably enter the world with few (Mitchell, Keller, and Kedar-Cabelli, 1986; DeJong
preexisting concepts. Instead, they must acquire or and Mooney, 1986). These models suggest that the
form concepts from experiences (e.g., form a concept most important aspect of concept learning is to ex-
of dog from existing experiences with dogs). As de- plain why a given example is an instance of the con-
scribed below, the classical, probabilistic, and theory cept. Construction of the explanation is carried out by
views of concepts propose different ways in which causally connecting known concepts. For example,
concepts are acquired. suppose a computer is to learn a concept cup and it
According to the classical view, the process of con- already knows such concepts as liftable, handle, liquid
cept formation is one of discovering necessary and container, and stable. Seeing an object that can be lift-
sufficient attributes by observing which attributes ed, has a handle, contains liquid, and is stable, the sys-
occur in all members and only in members of the cate- tem uses its background knowledge to construct an
gory. Research associated with the classical view has explanation about why one can drink from this object.
been directed at investigating hypothesis testing Then it generalizes this explanation to develop its
strategies, with each hypothesis being a guess as to concept of cup.
which features are part of the definition (Levine, Learning by analogy is another form of theory or
1971). knowledge-driven learning in which a known similar
In contrast, according to the probabilistic view, concept is modified. For example, one can learn
concept learning occurs by averaging values of mem- about the internal structure of atoms by applying

ones knowledge of solar system (e.g., electrons re- Kruschke, J. K. (1992). ALCOVE: An exemplar-based connection-
volve around the nucleus as planets revolve around ist model of category learning. Psychological Review 99, 2244.
Levine, M. (1971). Hypothesis theory and non-learning despite
the sun [Gentner, 1989]). One can also discover new
ideal S-R reinforcement contingencies. Psychological Review
features through analogy or metaphors (e.g., given a 45, 626632.
smile is like a magnet, one can learn that a smile at- Malt, B. C., Sloman, S. A., Gennari, S., Shi, M., and Wang, Y.
tracts). (1999). Knowing versus naming: Similarity and the linguistic
categorization of artifacts. Journal of Memory and Language 40,
The Why of Concept Learning Markman, A. B., and Makin, V. S. (1998). Referential communica-
There are many reasons why humans have con- tion and category acquisition. Journal of Experimental Psycho-
logy: General 127, 331354.
cepts. They allow people to classify things (e.g., recog-
Medin, D. L. (1989). Concepts and conceptual structure. American
nize that something is snake) and to make important Psychologist 12, 1,4691,481.
predictions or inferences (e.g., that snake may be poi- Mitchell, T. M., Keller, R. M., and Kedar-Cabelli, S. T. (1986). Ex-
sonous). John Anderson (1990) has developed a theo- planation-based generalization: A unifying view. Machine
ry of concepts that emphasizes this prediction func- Learning 1, 4780.
tion. Other functions include explanation (e.g., the Murphy, G. L., and Medin, D. L. (1985). The role of theories in
concept introvert might help to explain why some per- conceptual coherence. Psychological Review 92, 289316.
Posner, M. L., and Keele, S. W. (1968). On the genesis of abstract
son did not attend a party), reasoning (deriving
ideas. Journal of Experimental Psychology 77, 353363.
knowledge from the stored information), communi- Rips, L. J. (1989). Similarity, typicality, and categorization. In S.
cation, and conceptual combination (e.g., from the Vosniadou and A. Ortony, eds., Similarity and analogical reason-
concepts glass and elephant one might construct the ing. Cambridge, UK: Cambridge University Press.
combined concept glass elephant). Many approaches to Ross, B. H. (1997). The use of categories affect classification. Jour-
concept learning have focused only on the classifica- nal of Memory and Language 37, 240267.
tion function. However, functions of concepts interact Schyns, P. G., Goldstone, R. L., and Thibaut, J. P. (1998). The de-
velopment of features in object concepts. Behavioral and Brain
such that it is important to study multiple functions
Sciences 21, 154.
together. For example, Brian Ross (1997) found that Smith, E. E., and Medin, D. L. (1981). Categories and concepts. Cam-
the diagnosis (classification) of a disease was impor- bridge, MA: Harvard University Press.
tantly influenced by features that were relevant to its Spelke, E. S. (1990). Principles of object perception. Cognitive Sci-
treatment (Malt et al., 1999; Markman and Makin, ence 14, 2956.
1998, for other examples of interactions). Research- Wisniewski, E. J., and Medin, D. L. (1994). On the interaction of
theory and data in concept learning. Cognitive Science 18, 221
ers have begun to appreciate and investigate the vari-
ety of functions that concepts have.
Douglas L. Medin
PERSPECTIVE Revised by Edward Wisniewski

Anderson, J. R. (1990). The adaptive characteristic of thought. Hills-
dale, NJ: Erlbaum.
Brooks, L. R. (1978). Non-analytic concept formation and memory
for instances. In E. Rosch and B. Lloyd, eds., Cognition and cat- CONDITIONING
egorization. Hillsdale, NJ: Erlbaum.
Bruner, J. S., Goodnow, J. J., and Austin, G. A. (1956). A study of
[A survey of the forms of learning called conditioning will
thinking. New York: Wiley. be found under CONDITIONING, CLASSICAL AND IN-
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: STRUMENTAL. The following articles discuss specific as-
MIT Press. pects of the classical conditioning paradigm in greater de-
DeJong, G. F., and Mooney, R. J. (1986). Explanation-based learn-
ing: An alternative view. Machine Learning 1, (2) 145176.
Elio, R., and Anderson, J. R. (1981). The effects of category gener- CLASSICAL CONDITIONING: BEHAVIORAL
alizations and instance similarity on schema abstraction. Jour-
nal of Experimental Psychology: Human Learning and Memory 7,
Galton, F. (1879). Composite portraits, made by combining those SCHEMES FOR HIGHER-ORDER FEATURES OF
of many different persons into a single, resultant figure. Jour- CLASSICAL
nal of the Anthropological Institute 8, 132144. NEURAL SUBSTRATES OF CLASSICAL
Gentner, D. (1989). The mechanisms of analogical learning. In S. CONDITIONING
Vosniadou and A. Ortony, eds., Similarity and analogical reason-
ing. Cambridge, UK: Cambridge University Press.
Keil, F. C. (1989). Concepts, kinds, and cognitive development. Cam- A particular type of instrumental conditioning is more close-
bridge, MA: MIT Press. ly examined under OPERANT BEHAVIOR.]

CONDITIONING, CELLULAR AND generally activated by the US or whether qualitatively

NETWORK SCHEMES FOR HIGHER- different associations are formed between the CSs.
The notion that the same associative forms of synaptic
plasticity are involved in both the formation of the
Experimental work on invertebrates and vertebrates CS1-US and CS2-CS1 associations has been corrobo-
has begun to allow the analysis of the neural mecha- rated by studies in rat fear conditioning that have
nisms underlying second-order conditioning, block- shown that the same manipulation that blocks first-
ing, and contingency. Here we look at the experimen- order conditioning also blocks second-order condi-
tal and theoretical data that provide a cellular- and tioning without blocking the conditioned response
network-level description of higher-order forms of (Gewirtz and Davis, 1997). It was shown that infusion
classical conditioning. of an NMDA antagonist (which blocks associative syn-
aptic plasticity) into the amygdala prevented second-
order conditioning. This finding supports the notion
Experimental Models of Higher-Order that second-order conditioning relies on the same
Forms of Classical Conditioning forms of synaptic plasticity as classical conditioning.
The development of neural analogs of habitua-
tion, sensitization, and classical conditioning have
greatly enhanced the understanding of the neural Blocking
mechanisms responsible for these forms of learning. Blocking, along with contingency, establishes that
A logical extension of these studies is to develop neu- pairing of a CS with a US is not sufficient to produce
ral analogs of higher-order forms of classical condi- conditioning. Blocking shows that it is important that
tioning. Research has shown that because many the CS contains novel information about the US. A
higher-order forms of conditioning are present in in- typical blocking protocol (Kamin, 1968) consists of
vertebrates, these preparations may lead to the estab- two phases. First CS1 is paired with the US. In Phase
lishment of neural analogs of higher-order condition- II a compound stimulus CS1+/CS2 is paired with the
ing. Moreover, recent research in vertebrates has also US. In the case of complete blocking, CS2 does not
begun to examine the neural mechanisms of higher- undergo any conditioning. The argument is that even
order conditioning. though it was explicitly paired with the US, CS2 does
not provide any information that was not already pre-
dicted by CS1+.
Second-Order Conditioning
Blocking has been described in many vertebrates,
In second-order conditioning, a CS produces a in the snail Limax (Sahley et al., 1982) and in bees
conditioned response not by direct pairing with the (Smith and Cobey, 1994). A critical question regard-
US but by pairing with another CS that has previously ing the mechanisms underlying blocking is, How does
been paired with the US. The training protocol for preconditioning of CS1 prevent or block the condi-
second-order conditioning proceeds in two phases. tioning of CS2, even though CS2 is paired with the
During Phase I, CS1 is paired with the US (resulting US? One study on eyeblink conditioning in rabbits
in CS1+). During Phase II, a novel CS2 is paired with has suggested that blocking relies on active inhibition
CS1+. The defining feature of second-order condi- of the US pathway (Kim et al., 1998). Eyeblink condi-
tioning is that a conditioned stimulus (CS1+) func- tioning relies on the cerebellum. It was shown that
tions as a reinforcing stimulus for the conditioning of complex spikes in Purkinje cells of the cerebellum
a second conditioned stimulus (CS2). (which represent activity in the inferior olive) are elic-
Second-order conditioning has been demonstrat- ited by the CS-US pairing in nave animals but not by
ed in various invertebrate (e.g., Menzel, 1981; Sahley CS-US presentations in trained animals. Since inferi-
et al., 1982) and vertebrate preparations (Rescorla, or olive activity seems to mediate the US, these
1980). It has also been observed in the gill-withdrawal studies suggest that the US pathway is actively inhibit-
reflex in a reduced preparation in Aplysia (Hawkins et ed during Phase II of a blocking protocol. The active
al., 1998). In this preparation the mantle organs are suppression of the US response seems to be mediated
isolated together with the abdominal ganglia. Tactile by the inhibition in the inferior olive produced by the
and electrical stimulation of the siphon and gill can cerebellar output (representing CS1+). Indeed,
be used as CSs and US. In this study extinction of CS1 blocking of this inhibition during Phase II of training
resulted in extinction of CS2, suggesting the forma- prevented blockingthat is, it allowed CS2 to under-
tion of a direct association between both conditioned go conditioning. This manipulation did not prevent
stimuli. Indeed an important question regarding the first-order conditioning to the CS1 in a separate
underlying mechanisms of second-order condition- group of control animals. These experiments suggest
ing is whether CS1+ essentially takes over the system that first-order conditioning not only results in the

production of a CR but also causes active inhibition Figure 1

of the US pathway; they also suggest that this nega-
tive-feedback loop is involved in blocking (see below).

Rescorla (1968) suggested that it is not just the
number of CS-US pairings but also the correlation be-
tween the CS and the US, or the ability of the CS to
predict the US, that determines the occurrence of
conditioning. He demonstrated that if one group of
animals was presented with ten CS-US pairings, and
one group was trained with ten extra USs in addition
to the ten CS-US pairings (i.e., the probability of a US
given a CS equals .5), the latter group displayed less
conditioning. Despite both groups having received an
equal number of CS-US pairings, the group that re-
ceived the additional US presentations exhibited less
In most invertebrate (Abramson and Bitterman,
1986; Farley, 1987) and vertebrate preparations in
Schematic circuit of a model which exhibits second-order
which it has been examined, contingency has been
conditioning and blocking. Conditioned stimuli (CS) activate
observed. However, it is important to distinguish be-
sensory neurons (SN). In the nave state the SNs make weak
tween the different protocols used to study contingen- synapses unto both the motor neuron (MN) and the facilitatory
cy. The extra US presentations can take place before neuron (FN). These synapses can be potentiated by activity-
(US preexposure), interspersed with, or after (US dependent plasticity: Synapses undergo potentiation if they are
postexposure) the CS-US pairings. In Aplysia it has coactive with the (FN). A key element of the model is that
been shown that extra US presentations interspersed plasticity is occurring in parallel at two sites: SNMN and
with training decreases conditioning (Hawkins et al., SNFN. It is this feature that allows a previously conditioned
1986). Understanding the mechanisms underlying stimulus to function as a US.
contingency requires a determination of whether
both US pre- and postexposures decrease condition-
ing. This is because both US pre-and postexposure ef- Consider two CS pathways, CS1 and CS2, each
fects are difficult to explain with the same mecha- represented by a sensory neuron (SN1 and SN2), a
nism. For example, Hawkins and Kandel (1984) US pathway represented modulatory or facilitatory
suggested that contingency could result from habitua- neuron (FN), and a single output unit that generates
tion of the US pathway; this phenomenon, however, both the CR and UR, represented by a motor neuron
would not account for postexposure effects. Thus, it (MN). An important aspect of the network is that the
is possible that what is often thought of as a single CS neurons not only connect to the MN but also to the
higher-order form of conditioning may actually re- FN (see Figure 1). During first-order conditioning in
flect multiple mechanisms. which the SN1 and FN are paired (for example with
a 100-ms interval), the facilitatory neuron will pro-
duce associative plasticity at both the SN1MN and
Theoretical Models of Higher-Order Forms SN1FN synapse.
of Conditioning
Some researchers have proposed numerous
mechanistic models that address both the cellular and Second-Order Conditioning
network mechanisms underlying higher-order forms It is easy to see how this circuit is intrinsically able
of conditioning (Hawkins and Kandel, 1984; Gluck to account for second-order conditioning. Since plas-
and Thompson, 1987; Buonomano et al., 1990). ticity is also occurring at the SN1FN synapse, CS1
Many of these models share common conceptual fea- can take over the FN and become a US. When CS2
tures that we will examine with respect to a model of and CS1 are paired, CS1 will activate the FN and re-
second-order conditioning and blocking originally sult in conditioning of CS2. In this model, second-
developed in the context of the Aplysia siphon- order conditioning results directly from associative
withdrawal reflex (Buonomano et al., 1990). synaptic plasticity occurring at two synapses. Plasticity

at the SN1MN synapse is responsible for the gener- complex forms of learning may differ from simple
ation of the conditioned response, whereas plasticity forms of learning in either the form(s) of cellular plas-
at the SN1FN synapses accounts for second-order ticity involved and/or in the circuitry. However, the
conditioning. One important prediction of this type studies described above indicate that when the same
of model is that it should be possible to block second- forms of synaptic plasticity responsible for simpler
order conditioning without interfering with first- forms of associative learning are embedded in more
order conditioning by specifically blocking plasticity complex networks, more complex forms of learning
at the SN1FN synapse. may emerge, including blocking and second-order


Like second-order conditioning, blocking could
arise from the ability of a previously conditioned MEMORY PROCESSING IN BEES; INVERTEBRATE
CS1+ (SN1+) to take control of the FN. As in sec- LEARNING: ASSOCIATIVE LEARNING IN LIMAX
ond-order conditioning, during Phase I of training,
the synaptic strength of SN1 would increase and be- Bibliography
come strong enough to activate the FN. There are two Buonomano, D. V., Baxter, D. A., and Byrne, J. H. (1990). Small
consequences of this CS1-induced activity in the FN networks of empirically derived adaptive elements simulate
higher-order features of classical conditioning. Neural Net-
that would contribute to blocking. Note that if SN1+
works 3, 507523.
effectively activates the FN (and activity in the FN un- Gewirtz, J. C., and Davis, M., (1997) Second-order fear condition-
dergoes rapid depression), during the Phase II of ing prevented by blocking NMDA receptors in amygdala. Na-
blocking in which CS1+/CS2 are paired with the US, ture 368, 471474.
two important things happen: First, the FN will be ac- Gluck, M. A., and Thompson, R. F. (1987). Modeling the neural
substrates of associative learning and memory, A computa-
tivated approximately simultaneously with the onset
tional approach. Psychological Review 94, 176191.
of CS1+/CS2, essentially resulting in a 0-ms intersti- Hawkins, R. D., Carew, T. J., and Kandel, E. R. (1986). Effects of
mulus interval that should not result in conditioning interstimulus interval and contingency on classical condition-
of CS2. Second, since the FN undergoes depression, ing of the Aplysia siphon withdrawal reflex. Journal of Neuro-
the output of the FN in response to a US that followed science 6, 1,6951,701.
Hawkins, R. D., Cohen, T. E., Greene, W., and Kandel, E. R.
CS1/CS2 would be small or zero and would not sup-
(1998). Classical conditioning, differential conditioning, and
port associative conditioning of SN2. One subtle as- second-order conditioning in the Aplysia gill-withdrawal re-
pect of blocking pertains to the question of extinction flex in a simplified mangle organ preparation. Behavioral
of CS1+: if CS2 does not undergo conditioning be- Neuroscience 112, 636645.
cause the US in ineffective, why doesnt CS1+ under- Hawkins, R. D., and Kandel, E. R. (1984). Is there a cell-biological
alphabet for simple forms of learning? Psychological Review 91,
go extinction? Plasticity in the Aplysia sensory neurons
predicts that the effective interstimulus interval func- Kim, J. J., Krupa, D. J., and Thompson, R. F. (1988). Inhibitory
tion of plasticity will be state-dependent. That is, cerebello-olivary projections and blocking effect in classical
while a nonconditioned SN may not exhibit plasticity conditioning. Science 279, 570573.
with a 0-ms ISI, a conditioned SN can exhibit plastici- Rescorla, R. A. (1968). Probability of shock in the presence and ab-
sence of CS in fear conditioning. Journal of Comparative and
ty at 0 ms and thus prevent extinction of CS1+ during
Physiological Psychology 66, 15.
Phase II of blocking. (1980). Pavlovian second-order conditioning, studies in associa-
Therefore, in this model, although CS2 would be tive learning. New York: Erlbaum.
Smith, B. H., and Boey, S. (1994). The olfactory memory of the
paired with the US during Phase II of training, prior
honeybee Apis mellifera II Blocking between odorants in bi-
conditioning of CS1 would block associative enhance- nary mixtures. Journal of Experimental Biology 195, 91108.
ment of SN2. Thus, blocking could be supported by
Jennifer L. Raymond
activity-dependent neuromodulation (the mechanism
John H. Byrne
of classical conditioning) in combination with accom-
Revised by Dean V. Buonomano
modation or synaptic depression.

The results of both empirical and theoretical CONDITIONING, CLASSICAL AND
studies indicate that at least two factors determine the
ability of neural circuits to support various features of
learning: the learning rules or forms of synaptic plas- Classical (Pavlovian) and instrumental (Thorndikian)
ticity and the connectivity of the circuitry in which conditioning are the two most widely employed para-
those forms of plasticity are embedded. Therefore, digms for studying simple, associative learning result-

ing from the organisms exposure to the temporal are obtained. Autoshaping consists of response-
conjunction of two or more events. The fully specified independent presentations of a lighted manipu-
classical conditioning paradigm consists of a set of op- landum (e.g., lighted key) as a CS and activation of a
erations involving an unconditioned stimulus (US) reli- food magazine as the US; the target response is con-
ably producing an unconditioned response (UR) and a tact with the manipulandum (e.g., key pecking). Key
conditioned stimulus (CS) initially shown not to produce pecking is not an instrumental response, nor is it a
a response resembling the UR. The CS and US are UR appearing in the constellation of URs to food in
then presented repeatedly to the organism in a speci- the mouth (Woodruff and Williams, 1976). Hence, ac-
fied order and temporal spacing, and a response sim- quisition of a response in an effector system not elicit-
ilar to the UR develops to the CS that is called the con- ed by the US qualifies autoshaping as a new associa-
ditioned response (CR); that is, CS-CR functions are tive learning paradigm.
obtained. Control over the temporal conjunction of Some discriminative approach procedures have
the CS, US, and UR makes classical conditioning been designated as Pavlovian simply because an ex-
preparations ideal vehicles for studying associative plicit cue (CS) is presented and food or water, desig-
learning because they can uniquely specify stimulus nated the US, is made available at a fixed time follow-
antecedents to the target response. Various temporal ing CS onset (e.g., Holland and Rescorla, 1975) and
arrangements of the CS and US give rise to different the approach behavior, by definition instrumental to
forms of classical conditioning (e.g., delay, trace, simul- receipt of the reinforcing event, has been erroneously
taneous). Classical conditioning is called classical re- designated a CR. At present, the preceding para-
ward conditioning if the US is a positive stimulus and digms are widely employed in the study of associative
classical defense conditioning if it is negative stimulus. learning, but whether they will converge with the em-
The positive or negative designation depends on the pirical laws of CS-CR paradigms has yet to be deter-
independent demonstration of the organisms per- mined systematically. The S-S and discriminative ap-
forming instrumental responses necessary to obtain proach paradigms lack the capability of CS-CR
the US or to remove itself from the US, respectively. paradigms to exercise absolute control over the tim-
What distinguishes classical from instrumental condi- ing and sequencing of stimulus events; and to identify
tioning is that presentation or omission of the US is the stimulus antecedents to the target response from
independent of CR occurrence; and the definition of the outset of training. In addition, with CS-IR and dis-
a CR is restricted to a target response selected from criminative approach procedures, the target response
among those effector systems elicited as URs by the is instrumentally conditioned. Consequently, these
US. Adherence to both components of the definition paradigms might be expected to be even less likely to
of classical conditioning avoids common confusions display convergence with the empirical laws of CS-CR
and ambiguities with other associative learning para- paradigms. In any event, despite the greater technical
digms commonly designated as classical condition- demands of measuring URs and CRs in CS-CR re-
ing. search, their methodological characteristics favor
their use in the study of associative learning.
Stimulus-Stimulus Paradigms
The designation classical conditioning has Biological Substrates
been applied to paradigms meeting only the require- The CS-CR paradigms are ideally suited for the
ment that CS and US be administered independent study of the biological substrates of associative learn-
of the target response, ignoring selection of the UR. ing because the target response is defined anatomi-
As a consequence, the term classical conditioning cally by a set of movements or secretions. The UCSs
has been extended from Russian physiologist Ivan Pe- elicitation of the UCR permits identification of the
trovich Pavlovs CS-CR to stimulusstimulus (S-S) target responses final common neural pathway(s)
conditioning paradigms involving principally condi- outside the conditioning situation and, thereby, af-
tioned stimulus-instrumental response (CS-IR) and fords the opportunity to observe changes in its activity
autoshaping procedures. The CS-IR paradigms in- from the start of conditioning (Thompson, 1976). In
clude conditioned suppression and other classical- contrast, S-S contingency and discriminative ap-
instrumental transfer procedures in which the stimulus- proach paradigms are inherently unsuitable for
stimulus pairings of classical conditioning are con- studying the biological basis of learning. First, in
ducted with a CS and a biologically significant event CS-IR paradigms, changes in the instrumental target
(e.g., shock) but without measurement of a UR or CR. response are not the consequences of its participation
The CS is then presented during ongoing instrumen- in the learning process. Rather, the changes are the
tal behavior and its facilitory or disruptive effect on result of interactions of hypothetical (unobserved)
responding is measured; therefore, CS-IR functions CRs with the CS that are governed by prior CS-US

pairings. As a consequence, any neural analysis of and US is necessary for CR acquisition; and respond-
learning that is directed at changes in the target re- ing produced by the unpaired control is nonassocia-
sponse is pointless. Second, since the target response tive, since the randomized sequencing of CSs and USs
in the discriminative approach paradigm is outcome- at exceedingly long, random intervals prevents any
defined, a wide variety of different body movements CS-US contiguity effects. However, associative theory
can yield the required outcome. Therefore, it is virtu- and its unpaired control methodology have been
ally impossible to identify a final common pathway for challenged by a contingency hypothesis (Prokasy, 1965;
the movements that make up the response. Rescorla, 1967), which asserts that associative learn-
ing in classical conditioning can be viewed as deter-
Control Methodology mined by the statistical relationship between the CS
The associative nature of a conditioning prepara- and US. The hypothesis assumes that if US probabili-
tion has come to be determined by the contiguous oc- ty is greater in the presence of the CS than in its ab-
currence of the CS and US and a set of control opera- sence, a positive contingency prevails and excitatory asso-
tions intended to estimate the contribution of other ciative effects would accrue to the CS; conversely, if
possible processes to responding. All response sys- US probability is higher in the absence than in the
tems show some level of baseline activity, often raised presence of the CS, the negative contingency would
by UCS presentations, which can produce an acciden- yield inhibitory associative effects. Moreover, the con-
tal coincidence of the CS and target response. More- tingency hypothesis assumes the unpaired controls
over, the likelihood of a target response to the CS perfectly negative contingency would lead the CS to
may be systematically affected by alpha responses, acquire inhibitory associative effects. Hence, Rescorla
which are URs to the CS in the same effector system (1967) proposed a truly random control to provide an
as the target response; and pseudo-conditioned and sen- associatively neutral condition for assessing excitatory
sitized responses established on the basis of prior US- and inhibitory conditioning.
alone presentations. A detailing of the latency, dura-
tion, amplitude, and course of habituation of the Rescorla (1967) specified the truly random con-
alpha response with a control group given CS-alone trol as involving independent programming of the CS
presentations can provide a basis for eliminating al- and US or equal US probabilities in the presence and
phas from consideration as a CR, since they are usual- absence of the CS. However, delineating pairing/
ly of a shorter latency than CRs. Hence, if a sufficient- unpairing cannot be determined a priori but only em-
ly long CS-US interval is employed, both alphas and pirically. CS-US pairing is specified by the CS-US in-
CRs can be observed in the interval and scored ac- tervals demonstrated to produce CR acquisition for a
cordingly (Gormezano, 1966; Gormezano et al., specific preparation, while explicitly unpaired de-
1983). notes the use of stimulus intervals outside the effec-
The reinstatement or augmentation of alphas to tive CS-US intervals. Consequently, in the absence of
the CS through US-alone or CS-US pairings is re- an empirically derived metric (i.e., effective CS-US
ferred to as sensitization. After eliminating alphas from conditioning intervals) to designate paired and un-
consideration, the contribution of pseudo-CRs to CR paired conditions, it is virtually impossible to pro-
measurement can be assessed by presentations of the gram an associatively neutral truly random control or
US one or more times prior to CS presentations. The predictable excitatory or inhibitory conditioning
procedure frequently results in responses to the CS, groups. Rescorla, seeking to validate the truly ran-
labeled pseudo-CRs, which are treated separately dom control, reported that its CS had no effect upon
from CRs because of their occurrence in the absence avoidance conditioning (Rescorla, 1966) or upon re-
of CS-US pairings. However, the US-alone procedure sponding in a CS-IR study where shock-US probabili-
precludes trial-by-trial assessment of pseudo-CRs for ty in the presence and absence of the CS were equal
comparison with CRs. Accordingly, a single unpaired (Rescorla, 1968). However, these findings were chal-
control procedure has evolved in which CS-alone and lenged by CS-IR studies revealing that trial number
US-alone trials are presented randomly the same and frequency of chance CS-US pairings under a truly
number of times as the paired CS-US group, but at random control could substantially affect (excitatory)
variable CS-US intervals exceeding those effective for conditioning (Gormezano and Kehoe, 1975). Subse-
CR acquisition. Under the unpaired control, re- quently, Rescorla (1972) disavowed the contingency
sponses on CS trials (excluding alphas) provide a hypothesis and truly random control and reverted to
summative measure of pseudo-CRs and baseline re- the use of the unpaired control. Nevertheless, the
sponses. truly random control is still widely employed despite
Use of the unpaired control is based on associa- the detailing of additional methodological limitations
tive assumptions that temporal contiguity of the CS (Papini and Bitterman, 1990; Wasserman, 1989).

Instrumental Conditioning when the experimental subject performs the target

response, the contingent event is received by both
Instrumental conditioning procedures are all
subjects. Therefore, both members of the pair receive
characterized by a contingent relationship between
the same number and temporal distribution of stimu-
the organisms response and a stimulus. Typically, if
lus events; the only difference is that the experimen-
the stimulus increases, decreases, or leaves unaffected
tal but not the control subject always receives the rein-
the probability of the response, it is identified as posi-
forcing event after execution of the target response,
tive, negative, or neutral, respectively. Although such
while the yoked partner receives the reinforcing event
labeling appears to be circular, Edward Thorndikes
independent of execution of the response.
(1913) characterizations of stimuli as satisfying
(positive) and annoying (negative) were not circular Thus, the yoked-control design appears to be ad-
because they were specified by behavior changes in- mirably suited to test the (null) hypothesis that the
dependent of the target response. Noncircularity can temporal relationship between the response and sub-
also be achieved by demonstrating transitivity of stim- sequent stimulus event is irrelevant to the observed
ulus effects on the target response to other (new) re- behavior change. Unfortunately, the design con-
sponses. founds within-subject sources of random error with
the treatment effect: Control of stimulus events by ex-
A positive or negative contingency between the perimental subjects can allow for systematic differ-
target response and reinforcing stimulus gives rise to ences in the number of experimental subjects that are
a variety of instrumental conditioning paradigms. more affected by the stimulus event than their yoked
The five most extensively studied are reward, punish- partners. The possibility of such confounding has
ment, omission, escape, and avoidance, which derive from rendered the results of yoked-control designs neces-
responses producing a positive (reward) or a negative sarily ambiguous. As a consequence, a means for as-
(punishment) stimulus; preventing a positive (omis- sessing the contribution of the third nonassociative
sion) or negative (avoidance) stimulus from occur- factor to instrumental conditioning has not yet been
ring; and terminating a negative stimulus (escape). achieved.
Woods (1974), employing a classification schema that
includes operant conditioning and presence or absence See also: OPERANT BEHAVIOR; PAVLOV, IVAN;
of a discriminative stimulus, enumerated sixteen in- THORNDIKE, EDWARD
strumental conditioning paradigms. However, de-
spite repeated attempts at conceptual clarification,
Coleman, S. R. (1981). Historical context and systematic functions
the operant remains devoid of causal stimulus ante- of the concept of the operant. Behaviorism 9, 207226.
cedents (Coleman, 1981) and, consequently, it cannot Gormezano, I. (1966). Classical conditioning. In J. B. Sidowski, ed.,
be employed to study associative learning. The oper- Experimental methods and instrumentation in psychology. New
ant is applicable only to the study of performance York: McGraw-Hill.
Gormezano, I., and Kehoe, E. J. (1975). Classical conditioning:
variables affecting postasymptotic or steady-state re-
Some methodological-conceptual issues. In W. K. Estes, ed.,
sponding. Aside from the study of discrimination Handbook of learning and cognitive processes, Vol. 2: Conditioning
learning processes, discriminative instrumental condi- and behavior theory. Hillsdale, NJ: Erlbaum.
tioning paradigms are widely employed to assess the Gormezano, I., Kehoe, E. J., and Marshall, B. S. (1983). Twenty
effects of concurrent classical conditioning of fear years of classical conditioning research with the rabbit. In J.
M. Sprague and A. N. Epstein, eds., Progress in psychobiology
or incentive motivation to the discriminative stimu-
and physiological psychology, Vol. 10. New York: Academic Press.
lus upon the instrumental target response. Holland, P. C., and Rescorla, R. A. (1975). Second-order condi-
tioning with food unconditioned stimulus. Journal of Compara-
tive and Physiological Psychology 88, 459467.
Control Methodology Papini, M. R., and Bitterman, M. E. (1990). The role of contingen-
cy in classical conditioning. Psychological Review 97, 396403.
Any occurrence of the target response without Prokasy, W. F. (1965). Classical eyelid conditioning. Experimenter
prior conjunction with the reinforcing stimulus is des- operations, task demands, and response shaping. In W. F.
ignated a nonassociative response attributable to 1. Prokasy, ed., Classical conditioning: A symposium. New York: Ap-
base rate; 2. independent presentations of the rein-
Rescorla, R. A. (1966). Predictability and number of pairings in
forcing stimulus; and 3. presentations of the reinforc- Pavlovian fear conditioning. Psychonomic Science 4, 383384.
ing stimulus independent of the target response. Im- (1967). Pavlovian conditioning and its proper control pro-
plementing controls for the first two factors are self- cedures. Psychological Review 74, 7180.
evident, and achieving a control for the third factor (1968). Probability of shock in the presence and absence of
CS in fear conditioning. Journal of Comparative and Physiologi-
has been essentially limited to the yoked-control design.
cal Psychology 66, 15.
In the design, pairs of subjects are selected and one (1972). Informational variables in Pavlovian conditioning.
of them is randomly designated the experimental and In G. H. Bower and J. T. Spence, eds., Psychology of Learning
the other the control subject. During conditioning, and Motivation. New York: Academic Press.

Thompson, R. F. (1976). The search for the engram. American Psy- CONSOLIDATION
chologist 31, 209227.
Thorndike, E. L. (1913). Educational psychology. New York: Teach- See: AMNESIA, ORGANIC; ELECTROCONVULSIVE
ers College, Columbia University. THERAPY AND MEMORY LOSS; KAMINS
Wasserman, E. A. (1989). Pavlovian conditioning: Is contiguity ir- BLOCKING EFFECT: NEURONAL SUBSTRATES;
relevant? American Psychologist 44, 1,5501,551. MEMORY CONSOLIDATION: MOLECULAR AND
Woodruff, G., and Williams, D. R. (1976). The associative relation CELLULAR PROCESSES; MEMORY
underlying autoshaping in the pigeon. Journal of the Experi-
mental Analysis of Behavior 26, 113.
Woods, P. J. (1974). A taxonomy of instrumental conditioning.
American Psychologist 29, 584597.

I. Gormezano
DECLARATIVE MEMORY experience. It also supports representations of rela-
tionships among various events, providing the larger
Memory is the process or processes by which the brain record of ones experience. It provides the critical
enables us to represent experience and permits expe- means for rapidly representing events, those one-
rience to shape us. Rather than a unitary capacity sup- time arbitrary or accidental concurrences of people,
ported by a single set of processes, however, there are places, and things, and the spatial, temporal, and in-
different forms of memory, supported by multiple, teractional relations among them.
functionally, and anatomically distinct memory sys-
tems. The form of memory upon which we seem to Moreover, declarative memory enables one to
depend most in the activities of everyday life and learn arbitrary, nonderivable associations through
about which we can most readily reflect is declarative experiencefor example, learning the names con-
memory. nected with peoples faces, or their addresses and
telephone numbers. Declarative memory thereby
provides for representations of relations beyond the
Declarative and Procedural Memory
province of events, encompassing the relations
There are various proposed taxonomies of mem- among the facts that constitute our knowledge of the
ory, each offering a different account of the divisions world. This point leads to a further critical distinc-
among the memory systems of the brain. Most such tion: between episodic memory, which contains auto-
accounts distinguish between declarative memory biographical records of personally experienced
and procedural memory (Cohen and Squire, 1980; events, and semantic memory, consisting of world
Cohen, 1984). Declarative memory supports the on- knowledge stored outside of personal contexts (Tulv-
demand accumulation, storage, and retrieval of new ing, 1972). As fundamentally relational, capturing the
data about facts and eventsthe information that we relations among many different elements of knowl-
capture from our experiences through our represen- edge, both episodic and semantic memory are sup-
tations of it. In contrast, procedural memory supports ported by the declarative memory system.
the shaping of behavioral repertoires acquired
through experience. Declarative memory differs from A second critical property of declarative memory
procedural memory in being a relational memory sys- is representational flexibility (Cohen, 1984; Cohen
tem. and Eichenbaum, 1993). Declarative memories can
be activated by all manner of external sensory or even
purely internal inputs, regardless of the current con-
The Nature of Declarative Memory text. And they can be accessed by any number of dif-
Declarative memory supports representations of ferent brain processors, not only the ones involved in
relationships among the constituent elements of an initially acquiring the memories. Once accessed, they


can be manipulated and used flexibly to guide perfor- Schacter, 1987; Richardson-Klavehn and Bjork,
mance under an enormous range of testing condi- 1988), in which performance depends on using the
tions, including those differing significantly from the test item to permit conscious recollection of some spe-
circumstances of original learning. In this manner, cific prior learning experience and then inspecting
declarative memory serves as the relational database the contents. A successful outcome requires recall of
on which so much of cognitive processing and behav- the relation between the items to be tested and the
ioral performance depends. Among the brain systems study list or study experience.
that access and manipulate the declarative-memory
database are the frontal-lobe systems that support The deficit in amnesia is evident in all manner of
cognitively mediated and consciously aware process- relations, whether verbal or nonverbal, spatial or
es, including conscious introspection and verbal re- nonspatial, or episodic or semantic. As regards the
ports of the contents of ones memories. last, hippocampal amnesia typically affects both per-
sonal and public events (Sagar, Cohen, Corkin, and
Growdon, 1985; Zola-Morgan, Cohen, and Squire,
Deficit of Declarative Memory in Amnesia 1984); it includes not only autobiographical but also
world knowledge. One example is the profound defi-
Amnesia is a devastating memory impairment fol- cit shown by the patient H.M. in learning new vocabu-
lowing damage to the hippocampal system. Patients lary (word-definition relations) that has entered the
with hippocampal amnesia typically have a combina- language since the onset of his amnesia (Gabrieli,
tion of anterograde amnesia, an impairment in ac- Cohen, and Corkin, 1988).
quiring new memory, and retrograde amnesia, loss of
memories preceding the trauma. The deficits seem Despite profound and pervasive impairment of
confined to the domain of declarative memory memory, amnesic patients show impressive preserved
(Cohen and Squire, 1980; Cohen, 1984; Ryan, Alth- learning and memory abilities. Such patients can
off, Whitlow, and Cohen, 2000). Thus, amnesic pa- learn motor, perceptual, and cognitive skills even
tients show marked impairment in tasks or situations though they are unable to remember the experiences
in which performance depends on learning the rela- during which they acquired the skills. For example,
tions among items, especially items associated only amnesic patients were able to learn to read mirror-
arbitrarily or accidentally. For example, such patients reversed text in training extended over several days;
have great difficulty in remembering the events of and they retained the skill after three months despite
daily life. The amnesic patient H.M., who has been marked impairment in recollecting the training expe-
studied fifty years, since undergoing a surgical resec- riences or recognizing the words on which they were
tion of medial temporal lobe structures (Scoville and actually trained (Cohen and Squire, 1980). Preserved
Milner, 1957; Corkin, 1984), exhibits marked impair- memory is characteristic of performance that can be
ment on various tests of memory for public events based on tuning of skills in particular domains, built
that occurred after the onset of amnesia and can bare- up of incremental improvements in performance with
ly recall any personal events since the time of his sur- each exposure, and expressed in a repetition of the
gery (Sagar, Cohen, Corkin, and Growdon, 1985). original learning situationsuccessful performance
in this case does not require the flexible, relational
Formal laboratory testing confirms the deficit in representations of declarative memory (Cohen, 1984;
memory for relations. Paired-associate learning is es- Schacter, 1987; Gabrieli, 1998; Eichenbaum and
pecially useful in diagnosing amnesia; in this proce- Cohen, 2001).
dure, in which one must learn a set of arbitrarily
paired words, amnesic patients show severe impair-
ment, as they do on most list-based recall or recogni- Declarative Memory and Consciousness
tion-memory tasks, in which they are asked to commit
to memory a set of common words, faces, or visual ob- Declarative memory is critical for conscious intro-
jects presented in a study list and then to report (in spection and conscious recollection. But this system
recall tests) or to judge (in recognition tests) which does not mediate any particular aspect of conscious
items appeared on that particular study list. Because processing; rather, it provides the flexible access to
such common stimuli are familiar from a lifetime of information about relations among people, places,
previous experience, remembering of specific study objects, and actionsthe relational databasethat
items requires the linkage of their identity to this par- undergirds conscious recollection and introspective
ticular study list or learning experience, thereby call- reports. This view accounts for the amnesic deficits in
ing on declarative memory. memory for relations, even those do not enter into
the conscious awareness of normal subjects (Ryan, Al-
Amnesic patients are usually impaired on explicit thoff, Whitlow, and Cohen, 2000; Chun and Phelps,
or direct tests of memory (Graf and Schacter, 1985; 1999). It also underscores the affinities between

human and animal models of amnesia. Hippocampal memories are then flexibly accessible to various corti-
amnesia in rodents and nonhuman primates pro- cal processors in supporting cognitive processing and
duces the same dissociation among memory capaci- behavioral performance.
ties that are typical of human amnesia. Such animals
show impairments in learning and remembering spa- See also: AMNESIA, FUNCTIONAL; AMNESIA,
tial relations among environmental cues, configura- INFANTILE; AMNESIA, ORGANIC; AMNESIA,
tions of multiple perceptually independent cues, con- TRANSIENT GLOBAL; EPISODIC MEMORY; GUIDE
textual or conditional relations, and comparisons TO THE ANATOMY OF THE BRAIN; SEMANTIC
among temporally discontinuous eventsall of which MEMORY: COGNITIVE ASPECTS; SEMANTIC
require a relational form of memory. Yet the same MEMORY: NEUROBIOLOGICAL PERSPECTIVE
animals can show normal learning and remembering
of a large variety of conditioning, discrimination, and Bibliography
skill tasks, none of which require a relational form of Chun, M. M., and Phelps, E. A. (1999). Memory deficits for implicit
memory but rather only gradual, incremental contextual information in amnesic subjects with hippocampal
changes in bias or reactivity to individual items with damage. Nature Neuroscience 2, 844847.
Cohen, N. J. (1984). Preserved learning capacity in amnesia: Evi-
repeated exposure. dence for multiple memory systems. In N. Butters and L. R.
Squire, eds., The neuropsychology of memory. New York: Guilford
Press, pp. 83103.
Brain Mechanisms of Declarative Memory Cohen, N. J., and Eichenbaum, H. (1993). Memory, amnesia and the
hippocampal system. Cambridge, MA: MIT Press.
The critical role of the hippocampal system in
Cohen, N. J., Ryan, J., Hunt, C., Romine, L., Wszalek, T., and
declarative memory is evident in the phenomenology Nash, C. (1999). Hippocampal system and declarative memo-
of amnesia. Neurophysiological and neuroimaging ry (relational) memory: Summarizing the data from function-
studies of the hippocampal system also demonstrate al neuroimaging studies. Hippocampus 9, 8398.
its association with memory for relations. Hippocam- Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re-
tention of a pattern-analyzing skill in amnesia: Dissociation of
pal neurons encode various relationships among sig-
knowing how and knowing that. Science 210, 207210.
nificant elements of an experience, firing preferen- Corkin, S. (1984). Lasting consequences of bilateral medial tempo-
tially for particular conjunctions of the elements in ral lobectomy: Clinical course and experimental findings in
studies of freely behaving rodents (Wood, Dudc- H.M. Seminars in Neurology 4, 249259.
henko, and Eichenbaum, 1999; Eichenbaum et al., Eichenbaum, H., and Cohen, N. J. (2001). From conditioning to con-
scious recollection: Memory systems of the brain. NY: Oxford Uni-
2000). In functional neuroimaging studies of hu-
versity Press.
mans, hippocampal system activation arises whenever Eichenbaum, H., Dudchenko, P., Wood, E., Shapiro, M., and
the task engages memory for the relations among Tanila, H. (1999). The hippocampus, memory, and place
items (Henke, Buck, Weber, and Wieser, 1997; Cohen cells: Is it spatial memory or a memory space? Neuron 23,
et al., 1999). 209226.
Gabrieli, J. D. E. (1998). Cognitive neuroscience of human memo-
Amnesia indicates that the hippocampal system ry. Annual Review of Psychology 49, 87115.
must interact with other brain systems to effect declar- Gabrieli, J. D. E., Cohen, N. J., and Corkin, S. (1988). The im-
ative memory. Retrograde amnesia in cases of hippo- paired learning of semantic knowledge following bilateral
medial temporal-lobe resection. Brain and Cognition, 7, 157
campal amnesia can extend backward over variable
lengths of time, but it is never total; the storage of Graf, P., and Schacter, D. L. (1985). Implicit and explicit memory
long-term memory is never completely lost. Hence for new associations in normal and amnesic subjects. Journal
the hippocampal system cannot be the repository, or of Experimental Psychology: Learning, Memory, and Cognition 11,
permanent storage site, of all long-term memory. In- 501518.
Henke, K., Buck, A., Weber, B., and Wieser, H. G. (1997). Human
stead, the reciprocal connections of the hippocampal
hippocampus establishes associations in memory. Hippocam-
system with all the higher-order cortical processors pus 7, 249256.
allow it to mediate storage in interaction with neocor- Richardson-Klavehn, A., and Bjork, R. A. (1988). Measures of
tical sites. After the various cortical processors identi- memory. Annual Review of Psychology 39, 475543.
fy the constituent elements of the event or experi- Ryan, J. D., Althoff, R. R., Whitlow, S., and Cohen, N. J. (2000).
Amnesia is a deficit in relational memory. Psychological Science
ence, the hippocampal system binds together the
11, 454461.
multiple elements into long-term declarative memory Sagar, H. J., Cohen, N. J., Corkin, S., and Growdon, J. M. (1985).
representations that capture the relations among the Dissociations among processes in remote memory. Annals of
elements, with the individual elements or attributes the New York Academy of Sciences 444, 533535.
represented in distributed fashion in the relevant cor- Schacter, D. L. (1987). Implicit memory: History and current sta-
tus. Journal of Experimental Psychology: Learning, Memory, and
tical processors. Thus, the interaction of the hippo-
Cognition 13, 501518.
campal system with neocortical processors and stor- Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
age sites mediates the relational memory binding that bilateral hippocampal lesions. Journal of Neurology Neurosur-
allows the formation of declarative memory. Such gery and Psychiatry 20, 1121.
108 DJ VU

Tulving, E. (1972). Episodic and semantic memory. In E. Tulving Explanations of Dj Vu

and W. Donaldson, eds., Organization of Memory. New York:
Academic Press, 382403. Interpretations of dj have varied from the
Wood, E. R., Dudchenko, P. A., and Eichenbaum, H. (1999). The parapsychological (a telepathic reversion to a previ-
global record of memory in hippocampal neuronal activity. ous lifetime) to the psychodynamic (the mind neutral-
Nature 397, 613616. izes an emotional situation by displacing it into the
Zola-Morgan, S., Cohen, N. J., and Squire, L. R. (1983). Recall of
remote episodic memory in amnesia. Neuropsychologia 21,
past). These theses are thoroughly discussed by Sno
487500. and Linszen (1990) and Neppe (1983). We will exam-
ine several possible scientific interpretations.
Neal J. Cohen
Neurological Explanations
For more than a century, researchers have sug-
gested that a dj vu experience reflects a neurologi-
cal dysfunction. Since dj vu is part of the preseizure
DJ VU aura in some TLEs, it seems reasonable that dj vu
in nonepileptics may result from small temporal lobe
Dj vu is a memory phenomenon widely experi- seizures. Another interpretation is that the dj vu ex-
enced by the general public and is often cited in the perience results from a momentary delay in neuronal
popular literature. While the experience was de- transmission from the perceptual organ to the higher
scribed in the scientific literature as early as the mid- processing centers of the brain. This slight (several
1800s, researchers used a variety of different terms to millisecond) increase in the normal time taken to
describe the experience (e.g., paramnesia) until the transmit the message because of a synaptic dysfunc-
middle of the twentieth century. A standard defini- tion may lead to a misinterpretation of the informa-
tion of dj vu used today was provided by Neppe tion as being old.
(1983, p. 3): any subjectively inappropriate impres-
Another neurological interpretation involves two
sion of familiarity of a present experience with an un-
pathways rather than one. In the visual system, infor-
defined past.
mation is received in the cortex first from the primary
Survey research indicates that between half to and then a secondary pathway. When the normally
two-thirds of individuals have had a dj vu experi- brief delay between the two tracks lengthens, the usu-
ence, although this estimate varies considerably (30 to ally seamless integration of the two messages is dis-
97 percent) across three dozen investigations. Of rupted and experienced as two separate messages. A
those who experience dj vu, nearly all have had variation on this position is that the primary rather
more than one, and most experience one or more than the secondary neuronal path is inordinately
each year. A dj vu experience is typically triggered slowed, resulting in a reversal of the normal sequence
by a visual scene, lasts for a few seconds, and is associ- of messages. We routinely interpret information from
ated with mild stress, anxiety, or fatigue. The primary the primary pathway as our initial perception, so
psychological reaction is surprise, and the time sense when the secondary pathway arrives slightly before
seems to slow down. the primary pathway, the information from the pri-
mary pathway feels familiar because a memory
The incidence of dj vu decreases systematically match already existsit was established moments be-
with age, and this is probably due to an increase in so- fore.
cietal awareness across recent years. For example,
Memory Explanations
Gallup and Newport (1991) found that belief in dj
vu nearly doubled between 1978 and 1990. There It is possible that the individual experienced the
may be a decrease in reports of dj vu with age be- present situation not directly but through a magazine,
cause older adults grew up in a time when belief in movie, TV show, or dream. Considerable research
dj vu was not as widely accepted as it is today. suggests that source confusions are routine and that
dj vu may reflect a match with a memory created
Dj vu appears to be more common in better- by media or imagination. Another memory perspec-
educated and better-traveled individuals; there is no tive is that the type of cognitive processing is being
evidence of gender differences. Dj vu also occurs in duplicated rather than the actual memory. Retrieval
the aura preceding a seizure in some temporal lobe success often depends upon the correspondence be-
epileptics (TLEs). This trend motivated research to tween the way information is processed during input
determine whether dj vu is a potential diagnostic and retrieval. If the cognitive processing of new infor-
tool in seizure activity, epilepsy, and brain pathology. mation follows a set of mental procedures similar to
Accumulating evidence, however, does not support those of a prior experience, an unexpected sense of
such speculation. familiarity may result.

Another memory perspective is that one aspect of coding and retrieval are simultaneously activated, this
the present setting is objectively familiar, but con- spurious familiarity results in dj vu.
scious recognition fails when the object appears in a Finally, we may have two different varieties of
different context. The familiarity elicited by the un- consciousness: One processes information from the
identified object is then overgeneralized to the entire environment, whereas the other monitors our inner,
setting. On a visit to a friends home for the first time, mental world. Dj vu may occur when normal con-
for example, the grandfather clock in the corner is sciousness is diminished by distraction, fatigue, or sei-
identical to one in your aunts home. The implicit fa- zure, forcing reliance on the internal consciousness
miliarity of this object is not connected to the old operating from internally generated images, result-
object but instead misattributed to the objectively new ing in a misreading of a new experience as old.
In summary, dj vu is experienced by most peo-
Finally, the present setting may evoke a sense of ple on an average of once per year. Its incidence de-
dj vu because of a familiar organization of the ele- creases with age, increases with education, and is
ments in a scene. It is not the grandfather clock in the more common under stress or fatigue. Likely expla-
living room of your friends new home that is familiar nations include neurological dysfunction (seizure,
but rather that the rooms layout: a sofa to the right synaptic slowdown), implicit familiarity of objects, di-
of the love seat, with a stairway to the left and a grand- vided attention followed by full perception, and/or al-
father clock against the back wall, the same as in your teration in the normal function of two cognitive pro-
aunts house. cesses.
Attentional Explanations Bibliography
An ongoing perceptual experience may occasion- Gallup, G. H., and Newport, F. (1991). Belief in paranormal phe-
ally be divided into separate perceptions through dis- nomena among adult Americans. The Skeptical Inquirer 15,
traction or inattention. In this explanation, first pro-
Mack, A., and Rock, I. (1998). Inattentional blindness. Cambridge,
posed by Titchener in 1924, a brief initial perception MA: MIT Press.
under diminished attention is followed immediately Neppe, V. M. (1983). The psychology of dj vu: Have I been here be-
by a second perception under full attention. The sec- fore? Johannesburg: Witwatersrand University Press.
ond impression matches the ignored first glance Sno, H. N., and Linszen, D. H. (1990). The dj vu experience: Re-
membrance of things past? American Journal of Psychiatry 147,
taken moments earlier, giving rise to the feeling that
the current experience duplicates something experi- Titchener, E. B. (1924). A beginners psychology. New York: Macmil-
enced before. As you walk into a new hotel lobby talk- lan
ing on your cell phone, you process the scene without
Alan S. Brown
full attention. When you hang up, you perceive the
situation with full awareness and get the feeling that
you have been here before. Recent research on inat-
tentional blindness by Mack and Rock (1998) suggests
that individuals often miss clearly visible objects if
they are focused on something else, even when the The term dementia describes a decline of previously
unattended object is directly in front of them. acquired intellectual, or cognitive, skills. Memory loss
is the primary symptom of dementia, but other cogni-
Dual-Processing Explanations tive symptoms exist as well. A diagnosis of dementia
Routine cognitive experience may often involve requires an impairment in memory and at least one
the operation of two separate but interactive process- other cognitive domain (American Psychiatric Associ-
es. While both processes usually operate in concert, ation, 1994). This can include an impairment in com-
occasionally they might shift out of synchrony, or one prehending or expressing language, sustaining atten-
process might occur in the absence of the other. For tion, orientation (knowing where one is and the date
example, retrieval and familiarity usually operate in and time), visual perception, visual construction, or
an independent but coordinated manner, with recall executive (planning and organizing) skills. In addi-
accompanied by familiarity. When retrieval is activat- tion to cognitive deficits, dementia is often associated
ed without familiarity, a familiar setting seems mo- with behavioral and/or psychiatric changes such as
mentarily unfamiliar (jamais vu). Conversely, when poor judgment (for example, spending money reck-
familiarity is activated in the absence of retrieval, dj lessly or dressing inappropriately), delusions (false
vu is the result. Another interpretation is that memo- beliefs), or hallucinations (seeing, feeling, or hearing
ry encoding and retrieval usually operate indepen- things that are not actually there). Disinhibition, ex-
dently of each other. We experience a new situation; emplified by the use of inappropriate language (i.e.,
then we encode it into memory. However, if both en- swearing), the telling of off-color jokes, or acting

overly familiar with others, can also occur. Mood Dementia, Delirium, Senility, or Mental
symptoms, such as depression, apathy, or increased Retardation?
irritability, can accompany dementia as well.
Dementia is diagnosed only in the absence of de-
Dementia is not a disease, but a syndrome that lirium. Like dementia, delirium is associated with
has many causes, including head injury, vitamin defi- memory loss, but it is also associated with disorienta-
ciency, hydrocephalus, epilepsy, depression, medica- tion, confusion, and alterations in consciousness that
tion effects, and toxic exposure. In addition, there are fluctuate throughout the day. Unlike dementia, delir-
a number of diseases that can cause dementia. These ium typically begins suddenly and is usually the result
include Alzheimers disease, vascular disease, HIV in- of a treatable cause such as illness or the effects of
fection, multiple sclerosis, Parkinsons disease, Hun-
tingtons disease, Lewy body disease, and others. De-
pending on the cause of dementia, its onset may be Dementia and advancing age often co-occur be-
gradual or sudden, and its course may be progressive, cause the biggest risk factor for the development of
stable, or reversible. dementia is advancing age. Dementia is not, however,
an inevitable consequence of aging. The term senility
Pathophysiology of Dementia has been incorrectly applied to memory impairment
that accompanies old age. Senility actually refers only
The pathology associated with dementia varies,
to the age of onset of dementia; the term senile demen-
depending on the cause of dementia. The term neu-
rodegenerative is used to describe dementia caused by tia is sometimes used to refer to dementia that occurs
diseases that lead to the progressive death of nerve after age sixty-five. The term presenile dementia, in
cells (neurons) in the brain. Some diseases are associ- contrast, signifies dementia whose onset is prior to
ated with the development of abnormalities within or age sixty-five.
around neurons that interfere with communication Mental retardation differs from dementia in that
between brain cells and/or cause cell death. For exam- it describes intellectual and adaptive functioning that
ple, the amyloid plaques (insoluble deposits of a type
has developed subnormally. Thus, functioning has
of protein found between neurons) and neurofibril-
not declined from a previously higher level, as is the
lary tangles (abnormal collections of twisted threads
case with dementia. It is possible, however, for some-
found within neurons) associated with Alzheimers
one with mental retardation to acquire dementia. The
disease (AD) are found primarily in the cerebral cor-
most common example of this occurs in people with
tex (outer covering of the brain), hippocampus, and
Downs syndrome, most of whom will develop AD by
entorhinal cortex. Parkinsons disease, in contrast, is
the time they are in their fifties (Mann, 1993).
associated with Lewy bodies (accumulations of neuro-
filamentous material located within neurons) in the
substantia nigra and locus ceruleus (Kish, Shannak,
and Horneykiewicz, 1988). Pick bodies (a common
Memory Functioning in Dementia
neuropathological finding in frontotemporal demen- Due to similarities in the clinical presentation of
tia) are concentrated primarily in the frontal and/or patients with disorders affecting primarily subcortical
temporal cortex of the brain (McKhann et al., 2001). structures, which differs from that observed in de-
Cerebrovascular disease causes dementia by dis- mentia caused by cortical damage, the terms cortical
rupting blood flow to brain cells and is not considered dementia and subcortical dementia are often used. A cor-
to be neurodegenerative. If blood vessels in the brain tical pattern of cognitive deficits is characterized by
are blocked (e.g., by cholesterol), blood flow is con- an inability to perform purposeful motor movements
stricted. Such an event, called ischemia, results in oxy- (apraxia), language disturbance (aphasia), memory
gen deprivation, which can cause neuron death. Vas- impairment (amnesia), and difficulty with object rec-
cular dementia is the term used to refer to dementia ognition (agnosia). The most common disease caus-
caused by multiple vascular accidents, such as small ing cortical dementia is AD. Subcortical dementia, in
strokes (in which case it is sometimes referred to as contrast, has many more causes, such as Huntingtons
multiinfarct dementia), or a single stroke. Whether disease (HD) or Parkinsons disease (PD), and is char-
numerous small vessel changes can cause dementia is acterized by slowed mentation (bradyphrenia), for-
unclear, but with sufficient accumulation these getfulness, motor abnormalities, impaired planning
changes have been associated with significant cogni- and judgment (so-called executive dysfunction),
tive changes (Boone et al., 1992). There are, of and mood changes such as depression or apathy
course, many other mechanisms leading to death of (Cummings and Benson, 1984). Memory perfor-
brain tissue and dementia, but these examples ex- mance also differs between cortical and subcortical
plain some of the ways dementia can be caused. dementia (see Table 1).

Short-Term Memory Table 1

Short-term, or working, memory refers to informa-
tion that is stored for only a few seconds, such as when
one repeats a list of numbers immediately after hear-
ing them. This skill does not reliably differentiate pa-
tients with cortical from those with subcortical de-
mentia, as both have been shown to perform as well
as normal control subjects in repeating digits for-
wards, and deficits in their ability to repeat digits in
reverse order (Calderon et al., 2001; Redondo Verge,
Brown, and Chacon Pena., 2001). Patients with Lewy
body dementia, however, are impaired on both por-
tions of this task (Calderon et al., 2001).

Long-Term Memory
Long-term, or secondary, memory refers to memory
for information that must be recalled after some peri-
od of time, such as after few minutes or after many
years. Semantic memory, a type of long-term memory
relating to meanings of words and their relationships,
is typically impaired in patients with cortical demen- dementia is not improved. Patients with subcortical
tia. Perhaps the most impressive example of this defi- dementia, however, benefit from cues, and might
cit is observed in patients with fronto-temporal de- even perform as well as normal healthy adults. This
mentia (FTD), who can have very severe language difference is thought to demonstrate that cortical de-
impairment, such as word-finding difficulty, inability mentia causes an inability to encode new information,
to appreciate syntax, and/or impaired language com- whereas subcortical dementia causes a deficit in re-
prehension, with relative sparing of other cognitive trieving information. Tests of implicit (or procedural)
abilities (Mesulam, 2000). The word-finding difficulty memory (memory that occurs without conscious aware-
observed in AD is also an example of the impaired se- ness, such as the learning of motor movements) elicit
mantic memory seen among patients with cortical de- the opposite pattern. That is, subcortical dementia
mentia. Semantic memory is not typically impaired in
produces impairment on tests of implicit memory,
patients with subcortical dementia.
such as a mirror-tracing task, whereas patients with
Another type of long-term memory, episodic mem- cortical dementia improve as much as normal healthy
ory, concerns memory for events. Patients with AD are adults with practice, even if they have no recollection
typically severely impaired in their ability to recall re- of having performed the task previously (Butters,
cent events, but are much better able to recall events Heindel, and Salmon, 1990).
that occurred many years ago (Dorrego et al., 1999).
Patients with HD, however, have been found to
exhibit a flat gradient of remote memory, with Recent Developments in Dementia
equal impairment for all previous years (Albert, But- Research
ters, and Brandt, 1981). Whether this is the case for
patients with other types of subcortical dementia New Drug Treatments
(e.g., PD) has not been consistently shown (Fama et In 1996, the Food and Drug Administration ap-
al., 2000). proved the first drug (tacrine) to treat the memory
Long-term memory has been further divided into impairment that results from AD. The drug works by
explicit and implicit memory. Explicit memory (some- increasing the amount of a particular chemical, called
times used interchangeably with the term declarative a neurotransmitter, in the brain. This neurotransmit-
memory) is typically more severely impaired in cortical ter, called acetylcholine, is related to memory func-
than subcortical dementia. Both types of dementia tioning, and is depleted in the brains of people with
cause an impairment in this ability to recall informa- AD. Since approval of the first drug to treat AD, three
tion (e.g. a list of words, a story, or simple line draw- new drugs with similar mechanisms of action have
ings) that was presented earlier. Provided with cues, also been approved. Another development in clinical
however, such as the category from which the words trials research for memory disorders has been the in-
belong, or a recognition (e.g., multiple choice or yes/ vestigation of over-the-counter herbs such as Ginkgo
no) paradigm, the performance of those with cortical biloba and the antioxidant vitamin E.

Refining Diagnostic Criteria for Types of Beck, C., Cody, M., Souder, E., Zhang, M., and Small, G. W. (2000).
Dementia Dementia diagnostic guidelines: Methodologies, results, and
implementation costs. Journal of the American Geriatrics Society
Because dementia-causing disorders can be diffi- 48, 1,1951,203.
cult to diagnose during life, and because of new devel- Boone, K. B., Miller, B. L., Lesser, I. M., Mehringer, C. M., Hill-
opments in neuropathological techniques, criteria for Guttierrez, E., Goldberg, M. A., and Berman, N. G. (1992).
Neuropsychological correlates of white-matter lesions in
dementing disorders are evolving. To achieve consis-
healthy elderly subjects: A threshold effect. Archives of Neurolo-
tency, experts in these disorders meet to establish a gy 49, 549554.
consensus regarding the features that distinguish a Butters, N., Heindel, W. C., and Salmon, D. P. (1990). Dissociation
particular disorder from others. From 1984 to 1999, of implicit memory in dementia: Neurological implications.
fourteen consensus guidelines concerning dementing Bulletin of the Psychonomic Society 28, 359366.
Calderon, J., Perry, R. J., Erzinclioglu, S. W., Berrios, G. E., Den-
disorders were published (Beck et al., 2000). Exam-
ing, T. R., and Hodges, J. R. (2001). Perception, attention,
ples of these include Lewy body dementia (McKeith and working memory are disproportionately impaired in de-
et al., 1997) and FTD (McKhann et al., 2001). mentia with Lewy bodies compared with Alzheimers disease.
Journal of Neurology, Neurosurgery, and Psychiatry 70, 157164.
Mild Cognitive Impairment Cummings, J. L., and Benson, D. F. (1984). Subcortical dementia:
Sometimes patients do not meet criteria for de- Review of an emerging concept. Archives of Neurology 41, 874
mentia because they have a mild impairment in mem- 879.
Dorrego, M. F., Sabe, L., Cuerva, A. G., Kuzis, G., Tiberti, C., Bol-
ory but in no other cognitive domain, and their every- ler, F., and Starkstein, S. E. (1999). Remote memory in Al-
day functioning is not impaired. Scientists have zheimers disease. Journal of Neuropsychiatry and Clinical Neuro-
introduced the term mild cognitive impairment (MCI) to sciences 11, 490497.
describe this syndrome, and in 2001 consensus Fama, R., Sullivan, E. V., Shear, P. K., Stein, M., Yesavage, J. A.,
criteria for MCI were published (Petersen et al., Tinklenberg, J. R., Pfefferbaum, A. (2000). Extent, pattern,
and correlates of remote memory impairment in Alzheimers
2001). This disorder has received a great deal of at- disease and Parkinsons disease. Neuropsychology 14, 265276.
tention because, in many cases, it is considered to be Kish, S. J., Shannak, K., and Horneykiewicz, O. (1988). Uneven
a harbinger of AD (Morris et al., 2001). Because of the pattern of dopamine loss in the striatum of patients with idio-
availability of new drug treatments for the memory pathic Parkinsons disease: Pathophysiologic and clinical im-
impairment caused by AD, there is a great deal of in- plications. New England Journal of Medicine 318, 876880.
Mann, D. M. A. (1993). Association between Alzheimer disease and
terest in diagnosing this disease before it has prog- Down syndrome: Neuropathological observations. In J. M.
ressed to dementia. Berg, H. Karlinsky, and A. J. Holland, eds., Alzheimer disease,
Down syndrome and their relationship. Oxford: Oxford Universi-
ty Press.
Conclusion McKeith, I. G. et al. (1997). Dementia with Lewy bodies: Trying to
define a disease. Neurology 47, 1,1131,124.
Burgeoning interest in dementia research in the
McKhann, G. M., Albert, M. S., Grossman, M., Miller, B., Dickson,
late twentieth century has led to better understanding D., and Trojanowski, J. Q. (2001). Clinical and pathological
of its neuropathological underpinnings, treatments diagnosis of frontotemporal dementia. Report of the work
for its memory impairments, and refinements in the group on frontotemporal dementia and Picks disease. Ar-
diagnostic criteria for the disorders causing it. The chives of Neurology 58, 1,8031,809.
Mesulam, M. (2000). Aging, Alzheimers disease, and dementia:
importance of early diagnosis of neurodegenerative
Clinical and neurobiological perspectives. In Principles of Be-
disorders has led to an increasing efforts to identify havioral and Cognitive Neurology, 2nd edition. New York: Ox-
very mild, objective memory problems. Thus, charac- ford University Press.
terization of memory functioning continues to be an Morris, J. C., Storandt, M., Miller, J. P., McKell, D. W., Price, J. L.,
essential aspect for the diagnosis of dementia syn- Rubin, E. H., and Berg, L. (2001). Mild cognitive impairment
represents early-stage Alzheimer disease. Archives of Neurology
58, 397405.
Petersen, R. C., Stevens, J. C., Banguli, M., Tangalos, E. G., Cum-
See also: ALZHEIMERS DISEASE: BEHAVIORAL mings, J. L., and DeKosky, S. (2001). Practice parameter:
ASPECTS; ALZHEIMERS DISEASE: HUMAN Early detection of dementia: Mild cognitive impairment (an
DISEASE AND THE GENETICALLY ENGINEERED evidenced-based review). Neurology 56, 1,1331,142.
ANIMAL MODELS; COGNITIVE ENHANCERS; Redondo Verge, L., Brown, R. G., and Chacon Pena, J. R. (2001).
PHARMACOLOGICAL TREATMENT OF MEMORY Executive dysfunction in Huntingtons disease. Reviews in
DEFICITS Neurology 32, 923929.

Jason Brandt
Bibliography Ralph H. B. Benedict
Albert, M. S., Butters, N., and Brandt, J. (1981). Patterns of remote Revised by Cynthia A. Munro
memory in amnesic and demented patients. Archives of Neurol-
ogy 38, 495500.
American Psychiatric Association (1994). Diagnostic and Statistical See also: APLYSIA: DEVELOPMENT OF PROCESSES
Manual of Mental Disorders, 4th edition. Washington, DC: UNDERLYING LEARNING; CHILDREN,
American Psychiatric Association. DEVELOPMENT OF MEMORY IN; EARLY

EXPERIENCE AND MEMORY; LANGUAGE dimension, with gradients forming an inverted V

LEARNING; SCHOOL LEARNING shape (Siegel, 1972). Latent inhibition refers to re-
tarded classical conditioning as a result of preexpo-
sure to the CS. Inverted-V-shape generalization gra-
dients have also been observed with tonal stimuli
DIENCEPHALON trained with Pavlovs conditioned inhibition proce-
See: AMNESIA, ORGANIC dure (Mis, Lumia, and Moore, 1972). Inverted-V gen-
eralization gradients have been observed in pigeon
operant tasks (Hearst, 1969).

DISCRIMINATION AND Paradigms for Occasion Setting

Occasion setters are stimulus features, such as the
The decade of the 1990s witnessed acceleration in the presence of a light or tone, that serve as discrimina-
convergence of theoretical and experimental studies tive stimuli. For example, the presence of a light
of discrimination and generalization from the do- might signal that operant responses will be rein-
mains of classical conditioning and instrumental (op- forced. The absence of this light would signal that op-
erant) learning. Classical conditioning refers to the es- erant responses are not reinforced. In general, a
tablishment of behavioral adaptations (conditioned feature-positive paradigm is one in which the OS
responses; CRs) by the methods of Pavlov. Instrumen- signals reinforcement; a feature-negative paradigm is
tal learning is a general term for goal-seeking behav- one in which the OS signals the absence of reinforce-
ior, and operant conditioning refers to reinforcement ment. The presence of a light or tone might signal re-
learning by the methods of Skinner. The term discrim- inforcement, whereas its absence signals nonreinfor-
ination refers to the capacity of organisms to learn dif- cement. In classical conditioning a feature-positive
ferent modes of behavior depending on signals or occasion-setting task would involve adding a feature
cues from the environment about the imminence or to the CS. For example, if the CS is a tone, the addi-
accessibility of reinforcement. Generalization refers to tion of a light sets the occasion for reinforcement,
stimulus generalization, the capacity for signals or whereas the tone alone would not signal reinforce-
cues that are different from those used for establish- ment. A feature-negative task would be one in which
ing learned behavior to evoke this behavior. Stimulus the light, instead of signaling reinforcement when
generalization in classical conditioning refers to the ca- presented with the tone, would signal its absences.
pacity of a stimulus other than the conditioned stimu- If the occasion setter overlaps the CS and signals
lus to evoke a CR. In operant conditioning, one set the reinforcement, it can result in a more robust
of stimuli, an occasion setter (OS), might evoke the CR than would otherwise be the case. By contrast,
behavior controlled by another OS, depending on if OS overlaps the CS and signals the absence of
their shared features or similarity. Skinner coined the the US, it can inhibit the CR. This feature-negative
term occasion setter to refer to signals or cues that pre- discrimination recognizing the absence of an OS as a
dict reinforcement. In recent years, the term occasion signal of the absence of the US is a relatively difficult
setting has been extended to encompass both classical discrimination for animals to master. In fact, a classi-
and operant forms of behavioral learning. As a conse- cal conditioning situation like this is called a condi-
quence of this mixing, the terminology and para- tioned inhibition paradigm, and research has shown that
digms used in occasion setting research borrow from this is relatively difficult discrimination for animals to
the two domains. The mixture of the two domains has master.
led to a healthy integration of methods and ideas
(Schmajuk and Holland, 1998).
The convergence of ideas about discriminations Pattern Learning
and generalization from classical and operant condi- The relative difficulty in learning feature-
tioning began during the late 1960s, when the princi- negative discrimination as compared with feature-
ples of stimulus control enunciated by operant- positive discrimination learning extends also to the
conditioning studies involving pigeons were found to paradigms of positive and negative patterning. Dem-
extend to eyeblink conditioning in rabbits (Moore, onstrations of patterning in both classical and oper-
1972). Specifically, generalization along an auditory ant conditioning tasks involve two signaling stimuli,
frequency dimension shares many of the same charac- A and B, which could be occasion setters, and three
teristics as visual wavelength generalization in pi- types of trials. In positive patterning, the trial types
geons. Conditioned stimulus preexposure (latent in- are AB+, A, and B. Animals readily learn to re-
hibition) also generalizes along an auditory frequency spond more vigorously to the reinforced stimulus

Figure 1 Perceptual Learning

Exposure to discriminative stimuli enhances sub-
sequent discrimination learning, an effect known as
perceptual learning (Hall, 1991). The phenomenon of
perceptual learning might seem to contradict the idea
that exposure to stimuli retards later learning, as in
latent inhibition. But latent inhibition can actually ac-
count for perceptual learning, if we assume that pre-
exposed discriminatory stimuli share features in com-
mon. Latent inhibition develops to shared and
unshared features alike, but the common features lose
associability more rapidly than the unshared features.
Animals learn to ignore the shared features, allowing
the unshared features to dominate subsequent dis-
crimination learning and thereby facilitating perfor-
mance (McLaren and Mackintosh, 2000).
Not only do the shared features undergo more la-
tent inhibition than the unshared features, the un-
shared features also can become mutually inhibitory
through mechanisms of Pavlovian conditioned inhibi-
tion, provided the two stimuli are alternated during
the preexposure phase (Dwyer, Bennett, and Mackin-
tosh, 2001). The unique aspect of one preexposed
stimulus can actively suppress associations with the
Percentage of conditioned eyeblink responses in rabbits to the unique aspect of the other preexposed stimulus, and
target CS (X) in the presence of feature positive (A) and feature vice versa. Thus, any tendency for one stimulus to
negative (B) occasion setters as a function of the AX and BX elicit a representation of the other is reduced, and
intervals (0, 5, 15, 45 seconds) used in training. this promotes perceptual learning.

Acquired Distinctiveness
compound, AB, and to inhibit responses to A and Acquired distinctiveness refers to enhanced discrim-
B, where (+) denotes a reinforced trial and () de- ination learning to stimuli or stimulus dimensions
notes a nonreinforced trial. In negative patterning, that had been used in a prior discrimination task. Un-
the trial types are AB, A+, and B+. In order to be- like perceptual learning, however, acquired distinc-
have appropriately in a negative-patterning task, the tiveness appears to entail more than latent inhibition
animal must somehow learn that responding to the of shared stimulus features and mutual inhibition of
stimulus compound AB must be suppressed. This can unique features. The additional ingredient is correla-
happen only if the compound stimulus has unique tion with reinforcement. The nature of this correla-
features that are subsets of neither A nor B (Pearce tion remains in question. George and Pearce (1999)
and Bouton, 2001). That is, the compound has a have argued that acquired distinctiveness stems from
configural component that is shared by neither A nor prior learning about the relevance of the stimuli for
B. solving problems based on the same class of reinforc-
ers. The relevance-to-reinforcement account of ac-
The general difficulty in learning a negative- quired distinctiveness does not imply that prior learn-
patterning discrimination suggests that this configu- ing about the lack of correlation of the stimuli for
ral component is often overshadowed by the noncon- reinforcement impedes discrimination learning.
figural aspects of A and B, each of which in isolation
drives a tendency to respond by virtue of their associ-
ation with reinforcement. At the same time, some Time Discrimination
negative-patterning tasks can be mastered compara- Temporal control of behavior is one of the signa-
tively easily, suggesting that the compound AB forms ture features of operant conditioning methods, and
a unique pattern that, although similar to A and B, instances of temporal discrimination and generaliza-
nevertheless is treated as a whole, thereby allowing tion for duration has been well documented in oper-
the animal to master the discrimination. ant conditioning tasks in animals and humans (e.g.,

Wearden and Bray, 2001). Similar instances of tem- Schmajuk, N. A., and Holland, P. C., eds. (1998). Occasion setting.
poral control of behavior occur in classical condition- Washington, DC: American Psychological Association.
Siegel, S. (1972). Latent inhibition and eyelid conditioning. In A.
ing (Gallistel and Gibbon, 2000; Kehoe and Macrae, H. Black and W. P. Prokasy, eds., Classical conditioning II: Cur-
2002). Conditioned-response timing depends on the rent research and theory, pp. 231247. New York: Appleton-
CS-US interval(s) used in training. Conditioned re- Century-Crofts.
sponses are timed so that they achieve maximal am- Weardon, J. H., and Bray, S. (2001). Scalar timing without refer-
plitude at the anticipated time(s) of the US, and this ence memory? Episodic temporal generalization and bisec-
tion in humans. Quarterly Journal of Experimental Psychology
occurs whether the CS is defined operationally as the 54B, 289309.
onset or the offset of a stimulus (e.g., Kehoe, White, N. E., Kehoe, E. J., Choi, J-S, and Moore, J. W. (2000). Coef-
Schreurs, Macrae, and Gormezano, 1995). In classical ficient of variation in timing of the classically conditioned eye-
conditioning the timing of conditioned responses be- blink in rabbits. Animal Learning and Behavior 28, 520524.
comes more variable as the CS-US interval increases David R. Thomas
(Gallistel and Gibbon, 2000; White, Kehoe, Choi, and Revised by John W. Moore
Moore, 2000). Temporal discriminative control of
classically conditioned responses also occur in occa-
sion setting (Kehoe, Palmer, Weiderman, and
Macrae, 2000; see Figure 1). DISTRIBUTED PRACTICE EFFECTS
See also: CLASSICAL CONDITIONING: BEHAVIORAL Learning and memory are generally improved by
PHENOMENA repetition. However, not all repetitions are equally
beneficial. The effectiveness of repetitions depends in
part on their temporal distribution. A piece of infor-
Dwyer, D. W., Bennett, C. H., and Mackintosh, N. J. (2001). Evi-
dence for inhibitory associations between the unique elements mation studied on several occasions widely spaced
of two compound flavors. Quarterly Journal of Experimental Psy- apart in time will be remembered better than a simi-
chology 54B, 97107. lar fact studied on several occasions close in time.
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition-
ing. Psychological Review 107, 289344. The advantage of distributed repetitions over
George, D. N., and Pearce, J. M. (1999). Acquired distinctiveness spaced repetitions has long been known. Hermann
is controlled by stimulus relevance and not correlation with Ebbinghaus discussed distributed practice effects in
reward. Journal of Experimental Psychology: Animal Behavior Pro- his classic 1885 monograph on memory. He noted
cesses 25, 363373. that with any considerable number of repetitions a
Hall, G. (1991). Perceptual and associative learning. Oxford: Claren-
don Press. suitable distribution of them over a space of time is
Hearst, E. (1969). Excitation, inhibition, and discrimination learn- decidedly more advantageous than the massing of
ing. In N. J. Mackintosh and W. K. Honig, eds., Fundamental them at a single time (p. 89). Similarly, Jost (1897)
issues in associative learning, pp. 141. Halifax: Dalhousie Uni- formulated the advantage of distributed over massed
versity Press. repetitions as one of his laws of memory. In subse-
Kehoe, E. J., and Macrae, M. (2002). Fundamental behavioral and
findings in classical conditioning. In J. W. Moore, ed., A
quent decades, distribution of repetition became an
neuroscientists guide to classical conditioning, pp. 171231. New important manipulation in the study of learning and
York: Springer. memory. Because many different procedures were
Kehoe, E. J., Palmer, N., Weiderman, G., and Macrae, M. (2000). used and many conflicting results were found, the
The effect of feature target intervals in conditioned discrimi- overall pattern was long unclear, and investigators,
nations on acquisition and expression of conditioned nictitat-
ing membrane and heart-rate responses in the rabbit. Animal
such as Underwood (1961), sometimes despaired of
Learning and Behavior 28, 8091. being able to find consistent advantages for distribut-
Kehoe, E. J., Schreurs, B. G., Macrae, M, and Gormezano, I. ed practice over massed practice.
(1995). Effects of modulating tone frequency, intensity, and
duration on the classically conditioned nictitating membrane
Among researchers on human memory, an im-
response. Psychobiology 23, 103115. portant breakthrough was the research of Arthur Mel-
McLaren, I. P. L., and Mackintosh, N. J. (2000). An elemental ton (1967), who used a procedure that became stan-
model of associative learning: Latent inhibition and perceptu- dard for many subsequent investigators. Participants
al learning. Animal Learning and Behavior 28, 211246. saw a list of words presented one at a time. Some
Mis, F. W., Lumia, A. W., and Moore, J. W. (1972). Inhibitory con-
trol of the classically conditioned nictitating membrane re- words were shown once on the list, while others were
sponse of the rabbit. Behavior Research Methods and Instrumen- shown twice. Of the words that were shown twice,
tation 4, 297299. some were repeated in massed fashion; that is, they
Moore, J. W. (1972). Stimulus control: Studies of auditory general- were presented twice in a row. Other words were re-
ization in rabbits. In A. H. Black and W. P. Prokasy, eds., Clas-
peated in spaced or distributed fashion; that is, they
sical conditioning II: Current research and theory, pp. 206230.
New York: Appleton-Century-Crofts. had their occurrences separated by one or more inter-
Pearce, J. M., and Bouton, M. E. (2001). Theories of associative vening words. After presentation of the list was com-
learning in animals. Annual Review of Psychology 52, 111139. plete, participants were asked to recall the items in

any order. Melton found that the probability of recall Study-Phase Retrieval Accounts
for repeated items increased as a positive function of
Many different theories have been used to ex-
the number of intervening items between presenta-
plain distributed-practice effects. Although the de-
tions. This advantage for distributed repetitions over
tails of these theories vary, contemporary accounts
massed repetitions is often called the spacing effect;
can generally be grouped into three separate ap-
the fact that memory for spaced items may improve
proaches. One such approach emphasizes the impor-
somewhat as the distribution between repetitions is
tance of study-phase retrieval (Braun and Rubin,
increased further is sometimes called the lag effect.
1998). This approach assumes that one way in which
Are massed presentations ever more effective repetition helps to improve memory is by reminding
than distributed repetitions? Glenberg (1976) pres- the learner of earlier encounters with the studied in-
ented evidence that massed presentations may be su- formation. For example, if a particular word occurs
perior if memory is only required for a very short in- twice on a list, the second presentation may remind
terval. When the memory test is given almost the learner of the first presentation. The learner may
immediately after the last presentation, massed repe- then think about the first presentation again, and this
titions may lead to superior memory than spaced rep- retrieval and added rehearsal would make the first
etitions. Evidently, the benefit one obtains from hav- presentation particularly memorable. To explain why
ing multiple study episodes right before the test can the distribution of repetitions affects memory, one
sometimes be greater than the advantage usually must add on the assumption that this reminding pro-
gained from distributed rehearsal. Therefore, cram- cess benefits memory only if sufficient time has passed
ming may be a reasonably effective study strategy if to guarantee that the first presentation is not already
one is sure that the test will occur as soon as the cram- being consciously rehearsed.
ming is finished. However, distributed practice con-
It is reasonable that an encounter with a stimulus
sistently results in better long-term retention.
may serve as a reminder of earlier encounters. How-
ever, at least one critical claim of this approach is un-
Importance of Distributed-Practice Effects supported. A study-phase retrieval account predicts
One reason why the effects of distributed practice that memory for the first occurrence of a repeated
on memory are seen by many researchers as impor- item is affected by distribution of practice. Although
tant is that they have wide generality. One can find it is difficult to distinguish between memory for the
these effects in many different subject populations, first occurrence and memory for the second occur-
including nonhuman animals (Davis, 1970), human rence of a repeated item, what little evidence re-
infants (Cornell, 1980), children (Toppino, 1991), searchers have suggests that it is memory for the sec-
and the elderly (Balota, Duchek, and Paullin, 1989; ond occurrence, not the first, that is affected by
Benjamin and Craik, 2001). These effects are found distribution of practice (Hintzman, Block, and Sum-
on all sorts of memory tests and on a wide variety of mers, 1973). Although a study-phase retrieval process
different materials. may contribute to effects of distributed practice, there
is little direct evidence in favor of this approach.
Distributed practice can also be used to improve
learning and retention of meaningful material. For
example, Rea and Modigliani (1985) showed that dis- Deficient-Processing Accounts
tributed practice facilitates memory for spelling and
An alternative approach to explaining distribut-
multiplication facts. Reder and Anderson (1982)
ed-practice effects is to claim that learners do not pro-
found that distributed practice leads to improved
cess the second occurrence of an item as fully when
memory for information from textbooks, relative to
it is repeated in massed fashion as when it is repeated
massed practice. Dempster (1987) showed that learn-
in spaced fashion. That is, the second occurrence of
ing of the definitions of uncommon English words
a massed repetition is not given as much attention or
benefited from distributed practice. Bahrick and
rehearsal as the second occurrence of a spaced repeti-
Phelps (1987) demonstrated that retention of Spanish
vocabulary words over an eight-year period was great-
ly affected by distribution of practice. In short, distri- Zechmeister and Shaughnessy (1980) suggested
bution of practice seems to be as important in the that deficient processing of massed items may reflect
mastery of classroom material as it is in the memori- rehearsal strategies on the part of learners. For exam-
zation of lists in a laboratory. Dempster (1988) called ple, in the case of people memorizing a list of words,
for educators to be more sensitive to the importance the participants do not just rehearse each item as it
of this variable for classroom instruction, as distribut- is presented; rather, they presumably divide their re-
ing practice may improve retention without requiring hearsal between the current item and previous items.
additional time and resources. The amount of rehearsal that a person devotes to an

item presumably depends on how difficult the item each occurrence and that distributed rehearsal in-
appears to be. Zechmeister and Shaughnessy found creases the probability that variability in study will
that participants overestimate the extent to which occur.
they would remember massed repetitions. Because
people may mistakenly believe that massed items will
be easy to remember, they may not devote as much Multiprocess Accounts
rehearsal and attention to them as they do to spaced It seems increasingly likely that there will be no
items. Indeed, when participants are asked to re- single explanation for why distributed practice im-
hearse aloud or to pace themselves through a list, proves memory. Rather, there are several reasons why
massed repetitions seem to receive less rehearsal than distributed repetitions are more effective than
distributed repetitions. massed repetitions, and all three of the major ap-
proaches (study-phase retrieval, deficient processing,
Another version of this approach (advocated by
encoding variability) may apply under some circum-
Hintzman, 1974) attributes distributed-practice ef-
stances. Greene (1989) and Russo, Parkin, Taylor,
fects to deficient processing of the second occurrence
and Wilks (1998) have offered multiprocess accounts
of massed repetitions but views this as a result of an
that combine these approaches. The nature of the
automatic, unconscious process, not as the result of a
subject population, the stimulus material, and the ex-
deliberate strategy. This type of account has the ad-
perimental procedure all seem to determine the na-
vantage of being applicable to subject populations,
ture of the processes that underlie the advantage for
such as human infants, that show distributed-practice
distributed practice.
effects but that presumably do not employ conscious
rehearsal strategies. Although the theoretical explanation for the ef-
fects of distributed practice is likely to be complex,
there is no question that these effects are powerful
Encoding-Variability Accounts and of wide generality. The fact that distributed prac-
Encoding-variability accounts of distributed- tice leads to superior retention than massed practice
processing effects assume that spacing between repe- needs to be taken into account by both experimenters
titions facilitates memory by increasing the likelihood and educators.
that each occurrence of a repeated item is stored in
a very different way in memory. This sort of approach See also: REPETITION AND LEARNING
may best be explained by way of an analogy. Imagine
that a business office is very disorganized, and that Bibliography
workers often have difficulty finding papers they Bahrick, H. P., and Phelps, E. (1987). Retention of Spanish vocabu-
need. If they have an important document that they lary over 8 years. Journal of Experimental Psychology: Learning,
want to be able to find at a later time, it would make Memory, and Cognition 13, 344349.
sense for them to make multiple copies of it. Howev- Balota, D. A., Duchek, J. M., and Paullin, R. (1989). Age-related
differences in the impact of spacing, lag, and retention inter-
er, it would be wise for them to keep each of the cop-
val. Psychology and Aging 4, 39.
ies in different places and filed in different ways; this Benjamin, A. S., and Craik, F. I. M. (2001). Parallel effects of aging
would increase the probability that they would find at and time pressure on memory for source: Evidence from the
least one copy when they need it. Encoding-variability spacing effect. Memory & Cognition 29, 691697.
accounts assume that distributed practice increases Braun, K., and Rubin, D. C. (1998). The spacing effect depends on
an encoding deficit, retrieval, and time in working memory:
the probability that different occurrences of a repeat-
Evidence from once-presented words. Memory 6, 3765.
ed item will be stored in memory in different ways, Cornell, E. H. (1980). Distributed study facilitates infants delayed
thereby increasing the probability that a person recognition accuracy. Memory & Cognition 8, 539542.
would be able to retrieve at least one occurrence. Davis, M. (1970). Effects of interstimulus interval length and vari-
ability on startle-response habituation in the rat. Journal of
Different theorists have approached encoding Comparative and Physiological Psychology 78, 260267.
variability in different ways. Landauer (1976) argued Dempster, F. N. (1987). Effects of variable encoding and spaced
that distribution of practice directly influenced the lo- presentations on vocabulary learning. Journal of Educational
Psychology 79, 162170.
cation of memories of specific occurrences. Other (1988). The spacing effect: A case study in the failure to
theorists, such as Gartman and Johnson (1972) and apply the results of psychological research. American Psycholo-
Glenberg (1979), have emphasized that distribution gist 43, 627634.
of practice may increase the probability that a repeat- Ebbinghaus, H. (1885; reprint 1964). Memory: A contribution to ex-
ed item will be interpreted or analyzed differently at perimental psychology. New York: Dover.
Gartman, L. F., and Johnson, N. F. (1972). Massed versus distribut-
each occurrence. What these approaches have in com- ed repetition of homographs: A test of the differential-
mon is the assumption that repeated items are re- encoding hypothesis. Journal of Verbal Learning and Verbal Be-
membered better if they are studied differently at havior 11, 801808.

Glenberg, A. M. (1976). Monotonic and nonmonotonic lag effects effects in the central nervous system, allowing subjects
in paired-associate and recognition memory paradigms. Jour- to be used as their own controls. The ability to evalu-
nal of Verbal Learning and Verbal Behavior 15, 115.
(1979). Component-levels theory of the effects of spacing
ate memory performance in both a drugged and an
of repetitions on recall and recognition. Memory & Cognition undrugged state is particularly useful in human
7, 95112. studies, where the number of subjects can be limited.
Greene, R. L. (1989). Spacing effects in memory: Evidence for a An additional advantage of this approach is that using
two-process account. Journal of Experimental Psychology: Learn- the same subjects as both the control and experimen-
ing, Memory, and Cognition 15, 371377.
Hintzman, D. L. (1974). Theoretical implications of the spacing ef-
tal groups (within-subjects design) is a much more
fect. In R. L. Solso, ed., Theories in cognitive psychology: The Loy- powerful way to detect small drug effects. Using a
ola symposium. Hillsdale, NJ: Erlbaum. within-subjects experimental design is particularly
Hintzman, D. L., Block, R. A., and Summers, J. J. (1973). Modality useful for evaluating memory, because there can be
tags and memory for repetitions: Locus of the spacing effect. substantial differences in baseline performance
Journal of Verbal Learning and Verbal Behavior 12, 229238.
Jost, A. (1897). Die Assoziationsfestigkeit in Abhangigkeit von der
among individuals.
Verteilung der Wiederholungen. Zeitschrift fur Psychologie 14, Studies of the effects of drugs on memory are
436472. valuable for several reasons. First, it is important to
Landauer, T. K. (1976). Memory without organization: Properties
of a model with random storage and undirected retrieval.
know whether a drug that is being used to relieve a
Cognitive Psychology 7, 495531. particular condition can affect (usually impair) mem-
Melton, A. W. (1967). Repetition and retrieval from memory. Sci- ory as a side effect (see Table 1). For example, agents
ence 158, 532. that block the effect of acetylcholine, a neurotran-
Rea, C. P., and Modigliani, V. (1985). The effect of expanded ver- smitter long known to modulate memory, are fre-
sus massed practice on the retention of multiplication facts
and spelling lists. Human Learning 4, 1118.
quently used as part of the treatment for the move-
Reder, L. M., and Anderson, J. R. (1982). Effects of spacing and ment disorders of Parkinsons disease. At high doses,
embellishment on memory for the main points of a text. Mem- anticholinergic drugs produce profound memory im-
ory & Cognition 10, 97102. pairments. Second, experimental studies using drugs
Russo, R., Parkin, A. J., Taylor, S. R., and Wilks, J. (1998). Revising with defined mechanisms of action have helped to de-
current two-process accounts of spacing effects in memory.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
fine the involvement of specific neurochemical sys-
tion 24, 161172. tems in memory. One example is the class of drugs
Toppino, T. C. (1991). The spacing effect in young childrens free called benzodiazepines. These drugs facilitate the ef-
recall: Support for automatic-process explanations. Memory & fects of the inhibitory neurotransmitter GABA
Cognition 19, 159167. (gamma-aminobutyric acid) and impair memory. Un-
Underwood, B. J. (1961). Ten years of massed practice on distrib-
uted practice. Psychological Review 4, 229247.
fortunately, not all agents that affect memory act dis-
Zechmeister, E. B., and Shaughnessy, J. J. (1980). When you know cretely. Ethanol (alcohol) impairs memory by an un-
that you know and when you think that you know but you known mechanism that appears to involve actions on
dont. Bulletin of the Psychonomic Society 15, 4144. several neurotransmitter systems and on nerve-cell
Arthur M. Glenberg membranes. Third, studies with drugs offer the possi-
Revised by Robert L. Greene bility of helping to better define different types of
memory. Fourth, drug studies offer the hope of devel-
oping treatments for diseases with symptoms of mem-
ory impairment, such as Alzheimers disease.
See: COGNITIVE ENHANCERS; DRUGS AND Characteristics of Memory
There are several distinct memory systems in the
brain. One useful classification scheme divides long-
term memory into declarative (explicit) and non-
declarative (implicit) categories. Declarative memory
involves the conscious recall of facts or events (e.g., re-
membering a name), whereas nondeclarative memo-
DRUGS AND MEMORY ry is not conscious and is usually expressed through
The psychopharmacological approach to the study of performance (e.g., riding a bicycle). Subtypes of
memory involves a systematic examination of the be- memory within each of these systems rely on different
havioral changes that occur following the administra- brain regions and therefore involve different sets of
tion of psychoactive drugs. This approach comple- neural connections and combinations of chemical
ments neuropsychological research in its neuro- neurotransmitters and second-messenger systems. In
biological approach to the study of learning and addition to distinctions based on neuroanatomy,
memory. Most psychoactive drugs produce reversible there are separate processes governing short-term

and long-term memory. This diversity implies that Table 1

there are a number of sites where drugs can modulate
memory and that different drugs should influence
different types of memory.
There are at least three components of memory:
acquisition, consolidation, and retention or retrieval.
Each can be affected by drugs. The most frequently
documented effects of drugs on learning and memory
pertain to acquisition. In such studies researchers ad-
minister the drug before training begins. Arousal and
attention influence acquisition, so sedatives usually
impair memory, whereas stimulants usually enhance
it. Mood and motivation also affect acquisition;
agents that reduce either usually impair memory. At
the most basic level, drugs can impair acquisition by
interfering with sensory perceptionblurring vision,
for example. Possible drug effects on peripheral func-
tions must be controlled for in experimental studies,
especially when animals are used as subjects.
Administering the agent after training helps to
disentangle the memory-altering properties of drugs
from their effects on sensory or motor functions. This
is the protocol used to study the effects of drugs on
memory consolidation. If the drug has been metaboli-
cally eliminated by the time memory is tested, then
posttraining treatments affect only consolidation, not
acquisition or retention. Studies of consolidation in
animals have provided valuable information about
the biochemical mechanisms of memory formation
and the intervals during which learning activates
these mechanisms.
The research on drug effects on retrieval is not as
extensive as that on acquisition or consolidation. Per- structures used only in scientific research. Most com-
haps the best-known example of drug-induced alter- pounds mentioned here either have a role as medi-
ations in retrieval comes from the literature on state- cines or are well known among scientists who study
dependency. State-dependent retrieval refers to the abil- the neuropharmacology of memory.
ity of subjects to retrieve information better when The most pernicious drug-induced impairment
they are in the same state as they were when the mate- of memory is the destruction of neurons in brain re-
rial was acquired. For example, if material was gions involved in memory formation. For example,
learned under the influence of a drug such as alcohol, domoic acid destroys neurons in the hippocampus,
retrieval under sober conditions is often worse than an area that is critical for the formation of long-term
if the subject is again intoxicated. Some such effects declarative memories. Other agents act by severely
hinge on the nature of the retrieval test, and not all
depleting levels of neurotransmitters in pathways that
results appear to depend on the state of the subject.
modulate memory. One example of this phenome-
non is the recreational drug Ecstasy (MDMA), which
Drugs That Impair Memory impairs memory by reducing serotonin levels. While
the effects of these types of treatments are long last-
Many chemical compounds are known to impair ing, the memory impacts of most drugs are reversible.
memory; only a general overview of these agents is
possible here. The initial question is whether a given Many drugs that impair memory act as neuro-
compound can be considered a drug. At one end of transmitter receptors. Usually they reduce the func-
the spectrum are commonly available agents that nor- tion of a particular neurotransmitter. For example,
mally have another use. For example, sniffing the or- scopolamine blocks the receptor for acetylcholine,
ganic solvents (e.g., toluene) in glue will impair mem- propranol blocks a class of receptors for norepineph-
ory. At the other extreme are arcane chemical rine, and dizoclipine (MK-801) blocks a subtype of

glutamate receptors. Sedative drugs are the excep- For example, picrotoxin reduces the inhibitory effect
tion: Benzodiazepines (e.g., diazepam [Valium]) and of GABA receptors, while amphetamine causes the re-
barbiturates enhance the action of GABA, the prima- lease of norepinephrine. Acetylcholinesterase inhibi-
ry inhibitory neurotransmitter in the brain, thus dis- tors increase the effect of acetylcholine and are used
rupting memory. to treat the memory deficits of Alzheimers disease.
All of these drugs are effective memory enhancers,
Long-term memory formation requires activation
but they all have have serious side effects.
of certain enzymes such as protein kinases (e.g, PKC,
PKA, CaMKII). This activation is controlled by several Drugs can influence many aspects of brain func-
different neurotransmitters and can be affected by tion, including learning and memory. Most drugs ad-
drugs that act at any of their receptors. Also, specific versely affect memory, although some can enhance it.
inhibitors of protein kinases impair memory. An im- There is an increasing need to develop safe and effec-
portant consequence of protein kinase activation is tive drugs to treat memory problems caused by aging
the initiation of new protein synthesis. Inhibitors of or disease.
protein synthesis (e.g., cycloheximide or anisomycin) Bibliography
can disrupt long-term memory formation. Castellano, C., Cabib, S., and Puglisi-Allegra, S. (1996). Psycho-
A given drug seldom uniformly impairs all types pharmacology of memory modulation: Evidence for multiple
interaction among neurotransmitters and hormones. Behav-
of memory. For example, ethanol and the ben- ioral Brain Research 77, 121.
zodiazepine drugs affect mainly long-term memory, Castellano, C., Cestari, V., and Ciamei, A. (2001). NMDA receptors
leaving short-term memory relatively unaffected. and learning and memory processes. Current Drug Targets 2,
These drugs also preferentially disrupt declarative 273283.
memory while leaving nondeclarative memory rela- Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memo-
ry: A review. Psychology Bulletin 96, 518559.
tively intact. Duka, T. T., Curran, H. V. H., Rusted, J. M. J., and Weingartner,
There has been considerable interest in the possi- H. J. H. (1996). Perspectives on cognitive psychopharma-
cology research. Behavioral Pharmacology 7, 401410.
bility of using drugs in normal subjects to model clini- Ellis, K. A., and Nathan, P. J. (2001). The pharmacology of human
cal amnesias such as Korsakoffs syndrome or Alz- working memory. International Journal of Neuropsychopharma-
heimers disease. Knowledge of the neurophar- cology 4, 299313.
macology of a drug that would model a particular am- Farr, S. A., Flood, J. F., and Morley, J. E. (2000). The effect of cho-
nesic syndrome might suggest useful treatment strat- linergic, GABAergic, serotonergic, and glutamatergic recep-
tor modulation on posttrial memory processing in the hippo-
egies. The specific impairments seen in normal sub- campus. Neurobiology of Learning and Memory 73, 150167.
jects treated with alcohol resemble those of patients Lister, R. G., and Weingartner, H. J. (1987). Neuropharmacologi-
with Korsakoffs syndrome, and it has been suggested cal strategies for understanding psychobiological determi-
that scopolamine mimics some of the features of the nants of cognition. Human Neurobiology 6, 119127.
cognitive impairments seen in dementia patients. McGaugh, J. L., and Izquierdo, I. (2000). The contribution of
pharmacology to research on the mechanisms of memory for-
However, no drug treatment has been described that mation. Trends in Pharmacological Science 21, 208210.
completely mimics the pattern of memory disruption Squire, L. R., and Zola, S. M. (1996). Structure and function of de-
induced by a specific disease. clarative and nondeclarative memory systems. Proceedings of
the National Academy of Sciences of the United States of America 93,
Drugs That Enhance Memory Sutherland, R. J., Hoesing, J. M., and Whishaw, I. Q. (1990). Do-
moic acid, an environmental toxin, produces hippocampal
More drugs hinder memory than help it, so there damage and severe memory impairment. Neuroscience Letters
is a lively interest in discovering memory-enhancing 120, 221223.
drugs, especially because many neurological diseases Richard G. Lister
impair memory. Also, some memory loss occurs in old Herbert J. Weingartner
age even in healthy people. Finally, nearly everyone Revised by Gregory M. Rose
would like to have a better memory. An important un-
answered question is whether enhancing normal
memory is a reasonable possibility.
In general, the agents that enhance memory act
in the opposite way from drugs that impair memory. See: LEARNING DISABILITIES
EARLY EXPERIENCE AND LEARNING The purpose of this entry is to illustrate the ef-
fects of alterations in early environments on learning
The brain of the developing organism is a unique and capacity in the adult organism. Several areas of inves-
dynamic system. During the prenatal and postnatal
tigation have demonstrated the importance of early
periods the brain differs dramatically from that of the
experience for later behavior, using a number of dif-
adult. For example, it contains more synapses early
ferent models to examine this issue. Thus, depriva-
in development than it does at any other stage in life
tion of visual experience early in development has
(Purves and Lichtman, 1980). Receptors for a number
been shown to markedly affect adult vision; this also
of neuropeptides (e.g., oxytocin) are found in higher
affects the animals behavior (Hyvarinen and Hyva-
concentrations early in development than later in life
rinen, 1979). There is an extensive research on the ef-
(Shapiro and Insel, 1989). In certain brain areas (e.g.,
fects of enriched environments on subsequent
the suprachiasmatic nucleus of the hypothalamus),
learning ability (Rosenzweig, 1984). This entry exam-
glucocorticoid receptors are found in high concentra-
ines the role of early handling as a model of infantile
tions only during early ontogeny (Van Eekelen et al.,
1987). These are but a few examples that attest to the stimulation on learning in the adult.
differences in the brain during development. For the In 1956 S. Levine and colleagues reported that
most part, scientists have not determined the func- neonatal manipulations have profound effects on
tional significance of these neuronal features of the later behavior. In essence, their study showed that
newborn brain. neonatal handling (i.e., removing rat pups from their
One of the critical aspects of the developing brain mother for a few minutes and returning them to their
is its plasticity. Both physiological and environmental mother), or electric shock during the same period,
stimuli have been shown to profoundly, and often markedly improved the animals capacity to learn a
permanently, influence the functional capacities of conditioned avoidance response when tested as
the organism (Levine and Mullins, 1966). Neonatal adults. This procedure involves placing the animal in
disturbance of the normal hormonal milieu leads to a two-compartment chamber. Both sides of the cham-
some of the best-known cases of permanent alter- ber contain grid floors that can be electrified. The an-
ations of function induced by early manipulations of imal is presented with a signal, the conditioned stimu-
physiological events. Neonatal exposure to hetero- lus, which is followed after a brief time by an electric
typical gonadal hormones, for example, results in a shock. The animal is required to cross from the elec-
permanently altered reproductive physiology and be- trified side of the chamber to the safe side. If it crosses
havior. Both over- and underexposure to thyroid hor- within the interval between the onset of the signal and
mone during early development cause changes in the the onset of shock, it avoids the shock (conditioned
brain that produce learning disabilities. avoidance). However, if the animal fails to cross dur-


ing this interval and the shock is delivered, this re- In order to address the issue of the effects of early
sponse is considered an escape response. handling on learning, V. H. Denenberg and J. R. C.
Researchers conducted a number of studies fol- Morton (1962) studied the effect of early handling on
lowing their initial observations that demonstrated the ability of rats to learn tasks that did not involve
these marked differences between handled and non- noxious stimuli. These investigators examined the in-
handled pups. In order to verify whether the effects teraction between early handling and environmental
of early handling were age-dependent, early-handled enrichment on the ability of adult rats to solve a
pups were compared with nonhandled pups and with Hebb-Williams maze. After weaning handled and
animals that were handled after weaning. Pups ma- nonhandled rats were reared in a neutral, restricted,
nipulated as infants were found to show avoidance or enriched environment. The animals were then re-
learning superior to that of nonhandled pups. Post- quired to solve a sequence of twelve test problems.
weaning-handled animals were more similar to non- The results indicated that the animals ability to learn
handled rats. the maze was affected by the postweaning, but not by
the pre-weaning, manipulation. Thus, rats reared in
Although experiments showed that early han- an enriched environment made significantly fewer er-
dling improved avoidance learning, the underlying rors than those reared in neutral or restricted envi-
causes of this improvement remained to be deter-
ronments. Based on these results, it was concluded
mined. One possible explanation for these findings is
that pre-weaning handling affected emotional pro-
that emotional reactivity was modified as a function
cesses but not learning ability.
of these early experiences. The more efficient avoid-
ance learning may therefore be a consequence of re- In 1972 R. Wong attempted to clarify the issue of
duced arousal levels. Increasing the shock levels dur- whether early handling directly affects associative
ing avoidance conditioning results in an inverted U- learning or whether the improvement in learning is
shaped function with regard to acquisition of the due mainly to differences in emotional reactivity. An
conditioned avoidance response. Thus, acquisition is experiment was conducted that presumably could dis-
improved by very low levels of shock but impaired by criminate these two processes. Thus, subjects were
very high levels. trained to reach a criterion of stable performance on
There are numerous reports that early handling a positive reinforcement task (food) and then pun-
reduces emotional reactivity (Whimbey and Denen- ished with an aversive stimulus (shock) for making the
berg, 1967). This reduced reactivity is demonstrable reinforced response. Comparisons could then be
using both behavioral and physiological indices. Ac- made in terms of the degree of response suppression
tivity levels and defecation in the open field differ be- following the presentation of the aversive event; and
tween handled and nonhandled subjects. Handled the rate of recovery after removal of the aversive stim-
animals explore more actively (i.e., show less freez- ulus. Wong argued that if handled animals showed
ing), defecate less in the open field, are less neopho- greater response suppression and a slower recovery
bic (Weinberg, Smotherman, and Levine, 1978), and than nonhandled animals, this would indicate a direct
are less reactive to human handling as adults (Ader, effect of handling on learning. If the contrary oc-
1965). One of the more sensitive physiological indices curred, it could be assumed that the primary influ-
of arousal is the activation of the hypothalamic- ence on the behavior was attributable to differences
pituitary-adrenal (HPA) system (Hennessy and Le- in emotionality. Animals were trained to alternate
vine, 1978). Following stress, nonhandled animals goal boxes in a T-maze to obtain food. Once the crite-
show plasma corticosterone (the primary glucocorti- rion had been reached, they were given a shock in the
coid secreted by the rodent) elevations that are both goal box where they had obtained positive reinforce-
larger and more persistent than those found in ani- ment (S+ box). Testing began one day after the pun-
mals handled during infancy (Levine et al., 1967). ished trial. The animals were placed in the maze and
received food only in the S+ box. During the acquisi-
These differences in emotionality are consistently
tion phase the handled group made more correct re-
found using a variety of different testing paradigms.
sponses (alternations) than the nonhandled group,
However, the effects of early handling on avoidance
indicating superior learning on a positive reinforce-
conditioning appear to vary when the parameters are
ment task. Following the shock exposure handled rats
different from those used in the original studies. Both
made fewer choices to the food reward box (S+) than
R. Ader (1965) and J. Weinberg and S. Levine (1977)
nonhandled animals, suggesting that handled ani-
failed to find differences in conditioned avoidance re-
mals had superior associative learning.
sponse, although differences in emotional reactivity
were clearly present. The question of whether early Decades later A. Tang (2001) investigated the
handling influences associative learning therefore role of exposure to novelty on hippocampal depen-
still remained unanswered. dent learning. Neonatal rats were exposed for three

minutes daily to a nonhome environment and com- shock; and 3. testing, during which latent inhibition
pared with littermate controls that remained in the was indexed by the animals suppression of licking
home cage for the period with the mother removed. during tone presentation. As in the previous experi-
The novelty exposed rats showed more rapid acquisi- ment, latent inhibition was observed in the handled
tion as juveniles and greater retention in adulthood males and females and in nonhandled females, but
in a spatial learning task. Further, these animals also not in the nonhandled males. Based on both of these
showed greater retention in an odor discrimination studies, the conclusion was that early handling exerts
task. a beneficial influence on learning capacity in the
adult animal.
In many of the studies described above, learning
was investigated using behavioral situations in which Only a very limited literature has attempted to
an aversive motivational component was present, examine the neural substrates of the early handling
making it difficult to dismiss entirely the effects of phenomenon. Michael Meaney and his colleagues
emotional reactivity on learning. Latent inhibition, a (1988) studied the long-term influence of early han-
behavioral paradigm that avoids some of these prob- dling on the neuroendocrine regulation of the HPA
lems, has been used to examine the relationship be- system. These researchers reported a long-term down
tween handling and learning. This paradigm was de- regulation of glucocorticoid receptors (GR) in the
scribed in the context of classical conditioning. Ivan hippocampus of nonhandled, but not of handled,
Pavlov was the first to demonstrate that repeated ex- aged animals. They further reported that spatial
posure to a conditioned stimulus (CS), prior to pair- learning, a hippocampus-dependent process, is sig-
ing this stimulus with an unconditioned stimulus nificantly improved in aged animals that had under-
(UCS), impairs the rate of conditioning that subse- gone early handling. However, the aged nonhandled
quently occurs (Pavlov, 1927). Thus, repeated expo- animals appeared to have suffered hippocampal cell
sure to a stimulus that is not followed by meaningful loss. The differences in spatial learning may thus be
consequences renders this stimulus ineffective for due to the prevention of this cell loss by early han-
subsequent learning. dling. However, given the results presented by Tang,
it would appear that the influence of early experi-
I. Weiner, Schnabel, Lubor, and Feldon (1985) ences on hippocampal dependent learning is evident
employed the latent inhibition paradigm to study the throughout the life span. These studies do demon-
question of early handling and learning. The experi- strate the long-term consequences of early handling
ment was conducted in two phases. During the first for at least one aspect of neural regulation. Although
phase (preexposure) members of one group were the implications of this down regulation of GRs for
placed in a shuttle box and presented with sixty five- learning have not been extensively investigated, there
second tones. Members of the second group were is some evidence that administering specific GR an-
placed in the shuttle box for an equivalent time with- tagonists interferes with the acquisition of a spatial
out exposure to the tones. In the second phase the an- learning task (Oitzl, and de Kloet, 1992). Other as-
imals were presented with one hundred tone (CS) and pects of the HPA system also seem to be involved in
shock (UCS) pairings and the number of conditioned learning and memory (van Wimersma Greidanus,
avoidance responses was recorded. The results 1982).
showed that nonpreexposed handled animals exhib-
ited better avoidance learning than nonhandled rats,
thus replicating earlier findings. However, pre- Conclusion
exposed handled animals performed more poorly There are many examples in the literature of
than non-pre-exposed handled rats. The findings long-term consequences of manipulations during in-
were to some extent sex-dependent: Whereas preex- fancy that affect learning and memory. However,
posed handled males and females and preexposed most of these studies have utilized toxic agents result-
nonhandled females exhibited the latent inhibition ing in permanent and irreversible morphological and
(i.e., performed more poorly than nonpreexposed physiological changes in the central nervous system
animals), the nonhandled males did not show any ef- that are later reflected as impairments in the adult or-
fect of preexposure on the conditioned avoidance re- ganisms ability to learn (Grimm, 1987). Environmen-
sponse. tal enrichment and early handling constitute exam-
In a further study, latent inhibition was investi- ples of subtle environmental manipulations that
gated using a conditioned emotional response (CER) cause permanent alterations in adult function that fa-
to test the influence of early handling (Weiner, Fel- cilitate rather than impair adult learning abilities.
don, and Ziv-Harris,1987). The CER procedure was
conducted in three phases: 1. pre-exposure; 2. acqui- See also: CHILDREN, DEVELOPMENT OF MEMORY IN;
sition, in which the pre-exposed tone was paired with EMOTION, MOOD, AND MEMORY

Bibliography Weiner, I., Schnabel, I., Lubow, R. E., and Feldon, J. (1985). The
Ader, R. (1965). Effects of early experience and differential hous- effects of early handling on latent inhibition in male and fe-
ing on behavior and susceptibility to gastric erosions in the male rats. Developmental Psychobiology 18, 291297.
rat. Journal of Comparative and Physiological Psychology 60, 233 Whimbey, A. E., and Denenberg, V. H. (1967). Experimental pro-
238. gramming of life histories: The factor structure underlying
experimentally created individual differences. Behavior 29,
Denenberg, V. H. (1969). Open-field behavior in the rat: What
does it mean? Annals of the New York Academy of Sciences 159,
Wong, R. (1972). Infantile handling and associative processes of
rats. British Journal of Psychology 63, 101108.
Denenberg, V. H., and Morton, J. R. C. (1962). Effects of prewean-
ing and postweaning manipulations upon problem-solving
Seymour Levine
behavior. Journal of Comparative and Physiological Psychology 55,
Deborah Suchecki
Grimm, V. E. (1987). Effects of teratogenic exposure on the devel-
oping brain: Research strategies and possible mechanisms.
Developmental and Pharmacology Therapeutics 10, 328345.
Hennessy, M. B., and Levine, S. (1978). Sensitive pituitary-adrenal
responsiveness to varying intensities of psychological stimula-
tion. Physiology and Behavior 21, 295297.
Hyvarinen, J., and Hyvarinen, L. (1979). Blindness and modifica- Hermann Ebbinghaus was the founder of the experi-
tion of association by early binocular deprivation in monkeys.
mental psychology of memory. He laid the founda-
Child Care and Health Development 5, 385387.
Levine, S., Chevalier, J. A., and Korchin, S. (1956). The effects of tion for the scientific study of memory in a mono-
early shock and handling on later avoidance. Journal of Person- graph titled ber das Gedchtnis (1885), translated into
ality 24, 475493. English in 1913 under the title Memory: A Contribution
Levine, S., Haltmeyer, G. C., Karas, G. G., and Denenberg, V. H. to Experimental Psychology.
(1967). Physiological and behavioral effects of infantile stimu-
lation. Physiology and Behavior 2, 5559.
Levine, S., and Mullins, R. F. (1966). Hormonal influences on
brain organization in infant rats. Science 152, 1,5851,592.
Meaney, M. J., Aitken, D. H., van Berkel, C., Bhatnagar, S., and Ebbinghaus was born on January 23, 1850, at
Sapolsky, R. M. (1988). Effects of neonatal handling on age- Barmen, near Bonn, Germany. His father was a well-
related impairments associated with the hippocampus. Science
to-do merchant. He studied languages and philoso-
239, 766768.
Oitzl, M. S., and de Kloet, E. R. (1992). Selective corticosteroid an- phy at the University of Bonn. He served in the army
tagonists modulate specific aspects of spatial orientation during the Franco-Prussian War of 18701871, and
learning. Behavior Neuroscience 106, 6271. upon returning to the university completed his doc-
Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Arenp. London: toral dissertation in 1873. He then spent some five
Oxford University Press. years traveling in France and England. He began his
Purves, D., and Lichtman, J. W. (1980). Elimination of synapses in
research on memory at Berlin in 1878, spending
the developing nervous system. Science 210, 153157.
Rosenzweig, M. R. (1984). Experience, memory, and the brain. more than a year on the initial set of experiments.
American Psychologist 39, 365376. Upon completing these studies he became a private
Shapiro, L. E., and Insel, T. R. (1989). Ontogeny of oxytocin recep- lecturer at the University of Berlin in 1880, and he
tors in the rat forebrain: A quantitative study. Synapse 4, 259 continued his studies of memory. He repeated many
266. of the original experiments from 18791880 in 1883
Tang, A. (2001) Neonatal exposure to a novel environment en-
hances hippocampal-dependent memory function during in-
1884 and added new ones. He published the report
fancy and adulthood. Learning and Memory 8, 257264. on both series in his 1885 monograph.
Van Eekelen, J. A. M., Rosenfeld, P., Levine, S., Westphal, H. M.,
and de Kloet, E. R. (1987). Post-natal disappearance of gluco- Ebbinghauss life after he published his epoch-
corticoid receptor immunoreactivity in the suprachiasmatic making study was active and productive. He was ap-
nucleus of the rat. Neuroscience Research Communications 1, pointed a professor at the University of Berlin in
129133. 1886, remaining there until 1894, when he moved to
van Wimersma Greidanus, T. B. (1982). Disturbed behavior and
the University of Breslau. He stayed at Breslau for
memory in the Brattleboro rat. Annals of the New York Academy
of Sciences 394, 655662.
eleven years and then accepted an appointment at the
Weinberg, J., and Levine, S. (1977). Early handling influences on University of Halle. Over the years he became a
behavioral and physiological responses during active avoid- prominent and respected member of the new scientif-
ance. Developmental Psychobiology 10, 161169. ic discipline of experimental psychology. A major
Weinberg, J., Smotherman, W. P., and Levine, S. (1978). Early han- source of his renown lay in his textbook of general
dling effects on neophobia and conditioned taste aversion.
psychology, Grundzge der Psychologie (1897), which
Physiology and Behavior 20, 589596.
Weiner, I., Feldon, J., and Ziv-Harris, D. (1987). Early handling became the most widely read psychology text in Ger-
and latent inhibition in the conditioned suppression para- many. Ebbinghaus died of pneumonia at Halle on
digm. Developmental Psychobiology 20, 233240. February 26, 1909.

Ebbinghauss Approach to Memory

Before Ebbinghaus, the study of memory consist-
ed of philosophical armchair speculation concerning
remembering and forgetting in everyday life, and
clinical observations of patients with memory disor-
ders. The philosophical approach of the day